1StatisticalProcessControl-AdvancedCourse中级制成统计控制课程0501001stQtr2ndQtr3rdQtr4thQtr第三部分2Pre-Control的用途及方法Purpose:–Asimplifiedapproachtodeterminingwhenaprocessneedsadjustmentinordertoproduce0defectsUse:–TomaintainimprovementsmadeonY(KPOV)untilmethodstocontrolYorX(KPIV)viaSPCorMistakeProofingaredeveloped.–Asamonitoringmethodforaprocessthatisinstatisticalcontrolandhasbeenjudgedcapable»ForinstanceyourteamhasidentifiedX’sandtheirrelationshiptoY’sandhaveestablishedoptimumsettings&tolerances–AsagoodscreeningmethodthatcanpinpointwhereXandRchartsshouldbeapplied.Note:thisapproachisnotachartingmethod;rather,itmonitorstheprocesswithoutregardtotrends.3Pre-Control&+/-6sDesignDesignRequirementActualProcessCapabilityLocationofPre-controlLinesCp1.00Cp1.25Cp1.50Cp1.75Cp2.00Cp1.33Cp1.66Cp2.00Cp2.33Cp2.66Nominal+/-1/4tolerancerangeNominal+/-1/5tolerancerangeNominal+/-1/6tolerancerangeNominal+/-1/7tolerancerangeNominal+/-1/8tolerancerange4USLLSL2s-2s1.5s-1.5sPre-ControlLineShrinkWithCpImprovementCp=1Cp=2Cp=42s-2s5Pre-Control方法描述PRINTTOLERANCE1/2PRINTTOLERANCEYellowRedYellowGreenRedGreenNOMINALPCLINEPCLINELSLUSL7%7%86%-3s3sCp=1Baselinecase:1/141/1412/14Consecutive2pointsatyellowzones:Probability:=4*(1/14)*(1/14)=4/196=2%Meandonotmove6Pre-Control方法描述PRINTTOLERANCE1/2PRINTTOLERANCEYellowRedYellowGreenRedGreenNOMINALPCLINEPCLINELSLUSL43%7%0%50%-3s3sCp=1Baselinecase:6/140/147/14Consecutive2pointsatyellowzones:Probability:=(6/14)*(6/14)=36/196=18.4%OnedotatRedzone:7%Meanmove:-1.5s-1.5s0%0/141/147Pre-Control方法描述PRINTTOLERANCEYellowRedYellowGreenRedGreenNOMINALPCLINEPCLINELSLUSL2.3%2.3%95.4%-4s4sCp=1.33Anothercase:1/501/5024/25Consecutive2pointsatyellowzones:Probability:=4*(1/50)*(1/50)=4/2500=0.2%-2s2s8Pre-Control方法描述PRINTTOLERANCEYellowRedYellowGreenRedGreenNOMINALPCLINEPCLINELSLUSL30.2%0%69.2%-4s4sCp=1.33Meanshift-1.5s1/501/5024/25Consecutive2pointsatyellowzones:Probability:=(30.2%)*(30.2%)=9.1%OnedotatRedzone:0.6%-2s2s-1.5s0%0.6%92.DetermineProcessCapabilitybytaking5consecutiveparts.Whenallfivefallwithinthegreenzone,theprocessisconsideredcapable,incontroland“qualified.”YellowRedRedYellowGreen-1.5s-3.0s+1.5s+3.0sm7%7%86%Pre-Control方法描述1022.46THEVISIONOFSIXSIGMA3.SamplingPlan:CommenceproductionandtaketwoconsecutivepartsonaperiodicbasisYellowRedYellowGreenRed.07.07.86Pre-Control方法描述11YellowRedRedYellowGreen-1.5s-3.0s+1.5s+3.0sm.07.07.86YellowRedRedYellowGreen-1.5s-3.0s+1.5s+3.0sm.07.07.86YellowRedRedYellowGreen-1.5s-3.0s+1.5s+3.0sm.07.07.864.规则:WhenapartfallsintheRedzone,eitherside:StoptheprocessandinvestigatethecauseofvariationWhentwoconsecutivepartsfallinoppositeYellowzonesStoptheprocessandinvestigatethecauseofvariationWhentwoconsecutivepartsfallinthesameYellowzoneStoptheprocessandverify/adjustprocesssettings方法描述-使用规则12WhenbothpartsfallintheGreenzone,oroneintheGreenzoneandoneintheYellowzoneContinuetoruntheprocess5.ForlowerCpk’s(e.g.Cpk1.2)adjustthesamplingfrequencybydividingtheaveragetimeperiodbetweentwostoppagesby6.YellowRedRedYellowGreen-1.5s-3.0s+1.5s+3.0sm.07.07.86•Pre-controlcanalsobeusedforattributesbyusingappropriatelycoloredgo-nogogauges.•Pre-controlcanalsobeappliedtovisualcharacteristicsbyassigningvisualstandardsforthePClines方法描述-使用规则13μ+sμ+3sASuggestedFormatofColorPre-ControlChartTransferthevariabledatatostickincoloredLogSheetAftermachineadjustmentorsetup,datashouldwith+/-syellowlineOnlyforMachineSetupPassμμ+2sμ-sμ-3sμ-2ss=(Xi-X)N2μ:Spec.Center14过程能力15能力决定6Sigma质量描述设计规格1350DefectivePPM1DefectivePPB6Sigma6SigmaLSLUSLTarget16能力確定及改善規範限度對控制限度規範限度-透過工程上的要求決定是否接受某種特性;又稱為规格限度控制限度-以控制图上的線為基礎,以確定過程是否穩定。控制限度是由過程的數據運算出來的,並不是按規範限度註:因為這原因,我們需要學習過程能力以確保有能力進行這些過程。17規範限度對控制限度規範限度-透過工程上的要求決定是否接受某種特性;又稱為allowableprocessspread(參考EngineeringVariation)控制限度-以controlchart上的線為基礎,以確定過程是否穩定。控制限度是由過程的數據運算出來的,並不是按規範限度(參考ShewhartVariations)註:規範限度與控制限度兩者間並無任何關係18能力研究的類別因應各規範的不同特質,便產生出不同類別的能力研究。主要的類別如下:双边规格目標相等於規範的平均數目標不相等於規範的平均數单边规格越小越好(只適用於規範上限)越大越好(只適用於規範下限)μLSLUSLμLSLUSLμUSLμLSL19確定過程力的須知(ShortTermCapability)穩定過程正態分佈準確規範限度可接受的測量系統變差20能力决定步骤1.计划机器能力研究2.根据已知的最佳参数设定机器并记录输入参量的值3.进行产品加工4.观察加工工艺并做笔记5.测量并记录输出响应参量6.检查数据的正态分布情况7.计算数据的平均值,标准和极差8.确定控制限和工艺能力指数Cp/Cpk9.采用XBAR-R控制图,计算本工艺的稳定度系数21决定制程能力的步骤22正態分佈-Z值為何使用正態分佈?透過運用平均數及標準差,正態分佈有助簡便計算/評定次品比率。例:在正態曲線圖下,Zvalue及其面積基本統計及統計概念必須知道:•平均值(xbar)•標準差(S)•USL,LSL%Def=Zu+ZLXSLSLUSLZ=X-XSZu=USL-XS次品率23=受控的制程不一定为有能力的工艺,由有能力的制程做出的产品不一定符合规格=sR/d2制程能力s由控制图得出Cpu=Cp=(Cpl+Cpu)/2Cp--表征工艺或设备的相对于允许规格的能力客家允许宽度实际工艺分布宽度USL-LSL6sCp=USL-X3sCpl=X-LSL3s^^^^24Pp--工艺潜在能力–该指数描述了相对于客家规格,实际的长期制程分布的宽窄PP==USL-LSL潜在的制程能力Pp:s=(X-Xi)N-12Cp与Pp的区别在于Sigma计算方式不同总变差规格宽度工艺宽度6s25CpPpDuetokickofftheoutliersincontrolchartCp与Pp的比较26双边规格Target=SpecificationsMeanCpk=Cp(1-k)K=X-μ(USL-LSL)/227双边规格TargetSpecificationsMeanProcessPerformanceIndex,Ppk–ThisperformanceindexaccountsthedeviationofprocesscenteringandvariationfromTarget;theformulaasbelow.USL-XorX-LSLUSL-XorX-LSL3sR/d23sR/d2==EstimationofSigmaforCpkandPpkareremainedunchanged.Ppk=Cpk=3s3s28USL-μ1-*USL-μorUSL-XX-μ3sR/d23sR/d2--Cpk=Cpk=1-*μ-Xμ-LSLμ-LSL3sR/d2-Cpk=X-LSL3sR/d2-Cpk=or单边规格: