AIAGPPAPSPC

©2010BobDoeringAllRightsReservedCorrectSPCPROCESSCONTROLFORPRECISIONMACHINING©2010BobDoeringAllRightsReserved©2010BobDoeringAllRightsReservedSTATISTICALPROCESSCONTROLAIAGPPAP4thEdition2.2.11.3AcceptanceCriteriaforInitialStudyTheorganizationshallusethefollowingasacceptancecriteriaforevaluatinginitialprocessstudyresultsforprocessesthatappearstable.ResultsIndex1.671.33≤Index≤1.67Index≤1.33InterpretationMeetsacceptancecriteriaMaybeacceptableDoesnotmeetacceptancecriteria©2010BobDoeringAllRightsReservedSTATISTICALPROCESSCONTROLAIAGPPAP4thEdition2.2.11.5ProcesseswithOne-SidedSpecificationsorNon-NormalDistributionsTheorganizationshalldeterminewiththeauthorizedcustomerrepresentativealternativeacceptancecriteriaforprocesseswithonesidedspecificationsornon-normaldistributions.©2010BobDoeringAllRightsReservedSTATISTICALPROCESSCONTROLWhenthisisnottrue,usingthisanalysismayresultinunreliableinformation.AIAGPPAP4thEdition2.2.11.5ProcesseswithOne-SidedSpecificationsorNon-NormalDistributionsNOTE:Theabovementionedacceptancecriteria(2.2.11.3)assumenormalityandatwo-sidedspecification(targetinthecenter).©2010BobDoeringAllRightsReservedSTATISTICALPROCESSCONTROLAIAGPPAP4thEdition2.2.11.5ProcesseswithOne-SidedSpecificationsorNon-NormalDistributionsNOTE(cont.):Thesealternateacceptancecriteriacouldrequiredifferenttypeofindexorsomemethodoftransformationofthedata.Thefocusshouldbeonunderstandingreasonsforthenon-normality(e.g.isitstableovertime?)andmanagingvariation.©2010BobDoeringAllRightsReservedCONTROLCHARTBASICSNORMALPROCESS:INCONTROLWITHCHANCEVARIATIONINORDERFORAPROCESSTOBENORMAL,ITSHOULDBEABLETOBESETATTHEMEAN,ANDWILLCONTINUETORANDOMLYVARYABOUTTHEMEANWITHOUTANYOPERATORINTERVENTION!©2010BobDoeringAllRightsReservedCONTROLCHARTBASICSEXAMPLESOFPROCESSESWITHNORMALVARIATIONCUTTINGGRASS©2010BobDoeringAllRightsReservedCONTROLCHARTBASICSEXAMPLESOFPROCESSESWITHNORMALVARIATIONAUTOMATEDBAKERY©2010BobDoeringAllRightsReservedWHATISCONTROL?Aprocessincontrolisintheidealstate100%conformingandpredictablemustremainstableovertimemustoperateinastableandconsistentmannermustbesetattheproperlevel(centered)thenaturalprocessspreadmustnotexceedtheproduct’sspecifiedtolerance(capable)©2010BobDoeringAllRightsReservedWHATARETHETYPESOFVARIATIONCAUSES?COMMONCAUSEInherentintheprocessAffectseverypartExamples:gravity,airpressure,toolwearSPECIALCAUSE“Assignable”DoesnotaffecteverypartExamples:toolbreakage,startup,changeofoperators©2010BobDoeringAllRightsReservedTRADITIONALSPCExpectsall“specialcauses”havebeeneliminatedExpectstheremainingvariationisrandom,withmostvariationclosetoacentralvalueSeekstofindtrendsfromwithinanotherwiserandomoutputtoactuponwhentheyoccur©2010BobDoeringAllRightsReservedTRADITIONALSPCTraditionalSPCislikelookingoutintospaceseekingsignsoflifeSearchforExtraterrestrialIntelligence(SETI)©2010BobDoeringAllRightsReservedTRADITIONALSPCTypically,youlistentorandomradiofrequencynoiselookingfora“pattern”–asapatternshowsintent©2010BobDoeringAllRightsReservedTRADITIONALSPCIfyoufindsuchatrend,thenyoucantakeactiontodetermineitsorigin.Thisismore“monitoring”than“control”Wow!signalwasastrongnarrowbandradiosignaldetectedbyDr.JerryR.EhmanonAugust15,1977,whileworkingonaSETIprojectatTheBigEarradiotelescopeofTheOhioStateUniversity©2010BobDoeringAllRightsReservedPROCESSDEFINITIONSDEFINITIONOFPRECISIONMACHININGAprocesswherematerialisremovedbyacuttingsurface–suchasgrinding,honing,turning,milling,etc.Theprocessmustbecontrolledinamannerthatallvariation(vibration,bearings,gageerror)isstatisticallyinsignificantexcepttoolwear.©2010BobDoeringAllRightsReservedCONTROLCHARTDEFINITIONSTOTALVARIANCEEQUATION2T=2ToolWear+2MeasurementError+2GageError+2Material+2Temperature+2Operator+2Other©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTIONGRINDINGEXAMPLE©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION“ProcessControlandEvaluationinthePresenceofSystematicAssignableCause”,AshokSarkarandSurajit,Pal,QualityEngineering,Volume10(2),1997-1998©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION©2010BobDoeringAllRightsReservedPROCESSDISTRIBUTION©2010BobDoeringAllRightsReservedCENTRALLIMITTHEOREMThecentrallimittheorem(CLT)sufficientlylargenumberofindependenteachwithfinitemeanandvariance,willbeapproximatelynormallydistributed(Rice1995).randomvariables©2010BobDoeringAllRightsReservedCENTRALLIMITTHEOREMPrecisionmachininghassufficientlylargenumberofdependenteachwithfinitemeanandvariance,willbenotbenormallydistributedThecentrallimittheoremdoesnotapply!non-randomvariablesofindependentrandomvariables©2010BobDoeringAllRightsReservedCENTRALLIMITTHEOREM©2010BobDoeringAllRightsReservedCENTRALLIMITTHEOREM©2010BobDoeringAllRightsReservedCONTROLCHARTFEATURESInterestingpoints:TheMEANhasnorealvalueincontrollingaprocesswiththeuniformdistribution“Runningtot