【文章內(nèi)容簡(jiǎn)介】
?ReproducibilityR0==ReproducibilitySTD?O=ReproducibilityAV=?o=?=(EV)2nrAV=(originAV)2=(5?3)=?o=AV/=10/=(d2*istakenfrom,n:Numbersofparts。r:Trials)July2023Shanghai MSA39Appendix1: d2*withmgJuly2023Shanghai MSA40Appendix2: A2,D3, D4 with mJuly2023Shanghai MSA41Repeatability and Reproducibility Guide? RangeMethod? AverageandRangeMethod(Xbar/RMethod)? AnalysisofVariance(ANOVA)MethodJuly2023Shanghai MSA42Range Method?Therangemethodisamodifiedvariablegagestudywhichwillprovideaquickapproximationofmeasurementvariability.?Thismethodwillprovideonlytheoverallpictureofthemeasurementsystem.Itdoes notdeposethevariabilityintoRR.?ItisatypicalusedasaquickchecktoverifythattheGRRhasnotchanged.?Thisapproachhasthepotentialtodetectanunacceptablemeasurementsystem80%ofthetimewithasamesamplesizeof5and90%ofthetimewithasamplesizeof10.July2023Shanghai MSA43Method?2appraisersand5partsforthestudy?Bothappraisersmeasureeachpartonce?TherangeforeachpartistheabsolutedifferencebetweenthemeasurementobtainedbyappraiserAandthemeasurementobtainedbyappraiserB?Thesumofrangeisfoundandtheaveragerangeiscalculated.(R)?Thetotalmeasurementvariability( whered2*isfoundinAppendixC,withm=2g=numberofparts)July2023Shanghai MSA44CaseParts AppraiserA AppraiserB Range(A, B)1 2 3 4 5 Conclusion: The measurement system is in need of improvement.July2023Shanghai MSA45Average and Range MethodUnliketheRangeMethod,thisapproachwillallowthemeasurementsystem’svariationtobedeposedintotwoseparateponents,repeatabilityandreproducibility,butnotheirinteraction.TheANOVAmethodcanbeusedtodeterminetheinteractionbetweenthegageandappraisers,ifsuchexists.If Repeatability is large pared to reproducibility, the reasons may be:? Theinstrumentneedsmaintenance.? Thegagemayneedtoberedesignedtobemorerigid.? Theclampingorlocationforgagingneedstobeimproved.? Thereisexcessivewithinpartvariation.If Reproducibility is large pared to repeatability, then possible causes could be:? Theappraiserneedstobebettertrainedinhowtouseandreadthegageinstrument.? Calibrationonthegagedialarenotclear.July2023Shanghai MSA46研究方法1.Obtainasampleofn5partsthatrepresenttheactualorexpectedrangeofprocessvariation.2.RefertotheappraisersasA,B,C,etc.andthenumber1throughnsothatthenumbersarenotvisibletotheappraisers.3.CalibratetheGage.4.LetappraiserAmeasurenpartsinarandomorderandentertheresultsinrowappraisersB,Cmeasurethesamenpartswithoutseeingeachother’sreadingsthenentertheresultsinrows6and11,respectively.5.Repeatthecycleusingadifferentrandomorderofmeasurement.Enterdatainrows2,7,thedataintheappropriatecolumn.Ifthreetrialsareneeded,repeatthecycleandenterthedatainrows3,813.6.Whenlargepartsizeorsimultaneousunavailabilityofpartsmakesitnecessary:(a)LetappraiserAmeasurethefirstpartandrecordthereadinginrow1.LetappraiserBmeasurethefirstandrecordthereadinginrow6.LetappraiserCmeasurethefirstandrecordthereadinginrow11.(b)LetappraiserArepeatreadingonthefirstpartandrecordthereadinginrow2,appraiserBrepeatreadingonthefirstpartandrecordthereadinginrow7,appraiserCrepeatreadingonthefirstpartandrecordthereadinginrow12, thenrepeatthiscycleandentertheresultsinrows3,8,and13,ifthreetrialsaretobeused.7.Analternativemethodmaybeusediftheappraisersareonthedifferentshift.:(a)LetappraiserAmeasureall10partsandenterthereadinginrow1mthenhaveappraiserArepeatthereadinginadifferentorderandentertheresultinrows2and3.DothesamewithappraisersBandC.July2023Shanghai MSA47記錄表格實(shí)例July2023Shanghai MSA48Calculation1.RangeCalculatetherangeforeachappraisersCalculatetheaveragerangeforeachappraisersTotalaveragerangeCalculatetheCLJuly16