TUV德国莱茵技术六西格码培训资料32MeasurementSystem

Chapter3.2MeasurementSystemsAnalysis测量系统分析2测量是科学的基础“Ioftensaythatwhenyoumeasurewhatyouarespeakingaboutandexpressitinnumbers,youknowsomethingaboutit.”LORDKELVINTheScienceofMeasurement3EffectsofMeasurementErrorAveragesVariabilitymmmtotalproductmeas.system=+MeasurementSystemBiasMeasurementSystemVariabilityAccuracyPrecisions2total=s2product+s2meassystem4测量误差平均值变差mmm总产品测量系统=+测量系统的偏差测量系统的变差准确度精确度s2total=s2产品+s2测量系统SourcesofMeasurementVariation'MeasurementVariation'HumidityCleanlinessVibrationLineVoltageVariationTemperatureFluctuationOperatorTechniqueStandardProceduresSufficientWorkTimeMaintenanceStandardCalibrationFrequencyOperatorTrainingEaseofDataEntryAlgorithmInstabiltyElectricalInstabilityWearMechanicalinstabilityGageEnvironmentWorkMethods6测量误差的原因'MeasurementVariation'HumidityCleanlinessVibrationLineVoltageVariationTemperatureFluctuationOperatorTechniqueStandardProceduresSufficientWorkTimeMaintenanceStandardCalibrationFrequencyOperatorTrainingEaseofDataEntry电性能不稳定磨损机械不稳定性量具环境测量方法计算不稳定取得数据的难易操作员培训校准频率量具维护标准足够的工作时间标准操作规程操作员技术湿度清洁程度震动线电压波动温度波动7看到的不一定真实PossibleSourcesofProcessVariationLong-termProcessVariationShort-termProcessVariationVariationw/isampleActualProcessVariationStabilityLinearityRepeatabilityAccuracyVariationduetogageVariationduetooperatorsMeasurementVariationObservedProcessVariationWewilllookat“repeatability”and“reproducibility”asthesearetheprimarycontributorstomeasurementerror.Reproducibility9过程变差剖析长期过程变差短期抽样产生的变差实际过程变差稳定性线性重复性准确度量具变差操作员造成的变差测量误差过程变差观测值“重复性”和“再现性”是测量误差的主要来源再现性过程变差AccuracyAccuracy—Doestheaverageofthemeasurementsdeviatefromthetruevalue?Truevalue:TheoreticallycorrectvalueNISTstandardsBiasDistancebetweenaveragevalueofallmeasurementsandtruevalueAmountgageisconsistentlyofftargetSystematicerrororoffset11准确度(Accuracy)准确度(Accuracy)—测量的平均值是否与真值吻合?真值(TrueValue):理论上正确的值国际度量衡标准偏倚（Bias)测量值的均值与真值的距离测量系统持续地偏离目标系统错误BIAS—Isthedifferencebetweentheobservedaverageofthemeasurementandthereferencevalue.Thereference-valueisthevaluethatservesasanagreed-uponreference.Thereferencevaluecanbedeterminedbyaveragingseveralmeasurementswithahigherlevel(e.g.,metrologylab)ofmeasuringequipment.ObservedAverageValueReferenceValueBIASDefinition13BIAS—测量结果的平均值与参考值的差异.参考值（reference-value）是一个预先认定的参考标准.该标准可用更高一级测量系统测量的平均值来确定(例如：高一级计量室)观测平均值参考值偏倚BIASX1=0.75mmX6=0.8mmX2=0.75mmX7=0.75mmX3=0.8mmX8=0.75mmX4=0.8mmX9=0.75mmX5=0.65mmX10=0.7mmOnePartMeasuredTenTimesbyOneAppraiserWhatelsedoyouneedtodetermineBIAS?ThereferenceValuedeterminedbythelayoutinspectionequipment(ensurethisequipmentwentthroughaGageR&R)is0.80mm.Theprocessvariationforthepartis0.70mm.=0.75Bias=0.75-0.8=-0.05%Bias=100[0.05/0.70]=7.1%Thismeans7.1%oftheprocessvariationisBIASXX=10BIASEXAMPLE:15X1=0.75mmX6=0.8mmX2=0.75mmX7=0.75mmX3=0.8mmX8=0.75mmX4=0.8mmX9=0.75mmX5=0.65mmX10=0.7mm同一操作者对同一工件测量10次如果参考标准是0.80mm.过程变差为0.70mm=0.75Bias=0.75-0.8=-0.05%Bias=100[0.05/0.70]=7.1%表明7.1%的过程变差是偏倚BIASXX=10偏倚BIAS实例:PrecisionTotalvariationinthemeasurementsystemMeasureofnaturalvariationofrepeatedmeasurementsTerms:RandomError,Spread,Test/RetesterrorRepeatabilityandReproducibilitysssMSGO222=+17测量系统总变差通过重复测量的方法测量到的过程自然变差代表名词：重复性（Repeatability）和再现性（Reproducibility）sssMSGO222=+精确度（Precision)Precision:RepeatabilityTheinherentvariabilityofthemeasurementsystemVariationinmeasurementsobtainedwithagagewhenusedseveraltimesbyoneoperatorwhilemeasuringacharacteristicononepart.EstimatedbythepooledstandarddeviationofthedistributionofrepeatedmeasurementsRepeatabilityislessthanthetotalvariationofthemeasurementsystemsGsGRd=2*19测量系统内在的变异性基于重复测量的数据，用分组后组内的标准偏差来估算小于测量系统的总变差sGsGRd=2*重复性指同一人使用同一测量工具对同一对象（产品）的同一特性进行多次测量中产生的变差，用于估计短期的变差MasterValue精确度：重复性Precision:ReproducibilityOperatorvariabilityofthemeasurementsystemVariationintheaverageofthemeasurementsmadebydifferentoperatorsusingthesamegagewhenmeasuringacharacteristicononepartEstimatedbythestandarddeviationofthedifferenceinaverages,basedonmeasurementstakenbydifferentoperatorsMustbeadjustedforgagevariationReproducibilityislessthanthetotalvariationofthemeasurementsystemsOsORd=2*21精确度：再现性测量系统中操作员产生的变异基于不同操作者的测量数据，按操作员分组，通过组平均值的差来估。应扣除量具的因素（组内变差）比测量系统总变差小sOsORd=2*InspectorAMasterValueInspectorBInspectorCInspectorAInspectorBInspectorC再现性指不同的人在对同种特性进行测量时产生的变差LinearityDifferenceintheaccuracyvaluesofagagethroughtheexpectedoperatingrangeofthegage504030201055453525155StandardTrialsR-Squared=0.981Y=0.934227+0.994959XRegressionPlot504030201055453525155StandardTrialsR-Squared=0.982Y=0.245295+0.99505XLinearityisNotGoodGoodLinearityBadLinearity23线性（Linearity）量具在不同测量范围的准确度和精确度的变化，当测量范围较宽时尤为要关注504030201055453525155StandardTrialsR-Squared=0.981Y=0.934227+0.994959XRegressionPlot504030201055453525155StandardTrialsR-Squared=0.982Y=0.245295+0.99505XLinearityisNotGood好的线性差的线性StabilityThedistributionofmeasurementsremainsconstantandpredictableovertimeforbothmeanandstandarddeviationTotalvariationinthemeasurementsobtainedwithagage,onthesamemasterormasterparts,whenmeasuringasinglecharacteristicoveranextendedtimeperiod.Evaluatedusingatrendchartor