机械专业外文翻译平面磨削中形位误差的改进型离散系统模型.docx

机械专业外文翻译平面磨削中形位误差的改进型离散系统模型.docx

《机械专业外文翻译平面磨削中形位误差的改进型离散系统模型.docx》由会员分享，可在线阅读，更多相关《机械专业外文翻译平面磨削中形位误差的改进型离散系统模型.docx（19页珍藏版）》请在冰豆网上搜索。

机械专业外文翻译平面磨削中形位误差的改进型离散系统模型

JournalofMaterialsProcessingTechnology,210（2010）1794-1804

平面磨削中形位误差的改进型离散系统模型

Y.Gao1,X.Huang1,Y.Zhang1

香港科技大学机械工程系

Keywords:

Surfacegrinding;Partialremoval;In-processsensing;Precisioncontrol;Model;Formerror

Abstract:

Grindingremainsasoneoffewchoicesbeingabletomachineveryhardmaterialstodeliverultrahighprecisionathighmaterialremovalrateforefficiency.Effectivemodelsareneededforprecisioncontrolofthemachiningprocess.Sofar,fewstudiesonformerrorpredictionhavebeenreported.Machiningusuallybeginswithpartialremovalofworkpiecesurface.Withoutin-processsensing,systemparameterscouldnotbeaccuratelydeterminednorsurfaceforminformationthuspreventingusfrommodelingforprecisioncontrol.Inthisstudy,animproveddiscretesystemmodelandanin-processsensingtechniquehavebeenproposedtoaddressthepartialremovalandprecisioncontrolproblems.Modelsforpartialremoval,fullremoval,andsparkingoutconditionshavebeenestablished.Formerrorassessmentinthepartialremovalstagehasbeeninvestigated.Itisfoundthatthegrindingconstantisabletoreflectchangesinmachiningconditionsandisabletorepresentmachiningcapability.Alargergrindingconstantwillmeanareducedsizereduction.Furtherstudiesofthegrindingconstantarenecessary.Fortheaccurateestimationofthegrindingconstant,twoapproachesareproposed.Theiterativeapproachwasfoundmoresuitableandconvergent.Theproposedmodelsandin-processsensingtechniquewerevalidatedthroughexperimentaltestingintermsofworkpiecesurfaceformprofileyn（x,z0）,averagesizereduction

cn,surfaceformerrorEpvnandnormalgrindingforceFnn.Throughdetailedexaminationandcomparativestudies,theproposedmodelsandin-processsensingtechniqueofferedsignificantimprovementsrangingfromapproximately16.9%to23%,comparedwiththeexistingmodels.Exceptthegrindingforce,whichwasindirectlymeasuredthroughavoltagemeasurementapproach,theoverallrelativeerrorsbetweenthetheoreticalresultsandtheexperimentalresultsunderfullremovalconditionswerefoundrangedfrom2.08%to6.87%,indicatingtheimprovedprecisionpredictioncapabilitiesoftheproposedsystemmodel.Theexperimentalresultscanbeusedasasetofreferencesforfurtherstudiestoofferperformanceassessment,precisionprediction,processplanning,andprocessconditionmonitoringforthisimportantprecisionmachiningprocess.

1.Introduction

1.1.Modelforprecisioncontrol

Grindingisanabrasiveprecisionmachiningprocesswhichremainsasoneoffewchoicesbeingabletomachineveryhardmaterialstodeliverultrahighprecisionathighmaterialremovalrateforefficiency.Processofthekindiswidelyusedtoachievehighaccuracyforhighqualitymechanical,electrical,andopticalparts（KarpuschewskiandInasaki,2006）.Surfacegrindingisoneforprecisionmachiningofsurfaces.Toachievehigheraccuracyforqualitycontrol,itisessentialtodevelopeffectivemodelstorealizeprecisioncontrolofthemachiningprocess.

Forgrindingprocessmodeling,modelsofmultipleaspects（BałaszandKrólikowski,2007）,suchasmodelofgrain（Horng,2008;Mamalisetal.,2001）,modelofgrindingwheeltopography（Bigerelleetal.,2005;ZhouandXi,2002）,modelofheattrans-fer（Liaoetal.,2000）,modelofprocesskinematics

（Wecketal.,2001;Zhangetal.,2005）,modelofchipformation（GopalandRao,2004;Heckeretal.,2007）,modelofforce（HekmanandLiang,1999;JenkinsandKurfess,1999;Tangetal.,2008）,andmodelofpower（Nandiaetal.,2004）,havebeenexamined.Thegrainormaterialremovalmodel（Horng,2008）wasbasedonsurfaceasperitycontactmechanics.Theelastic–plasticeffectsinthewearmechanismwereconsideredtoberelatedtothedensityofabrasivegrains.Mamalisetal.（2001）proposedamodelforinteractionbetweenhardpolycrystallinematerialsandwheelgrainduringgrinding.Wornsurfacesofgrindingwheelmaybemodeledusingfractalfunctions（Bigerelleetal.,2005）.Aroughnesspredictionmodelforwheeltopography,wear,andgrindingkinematicswasestablishedbyZhouandXi（2002）.ThethermalmodelbyLiaoetal.（2000）involvedathermaleffectofgrainandworkpieceinterfaceandashearplanebetweenworkpieceandchip.Thetemperatureoftheworkpiecesurfaceinthegrindingzonecouldbepredicted.Adynamicbehaviormodelforthecylindricaltraversegrinding

processinthetimedomainwaspresentedbyWecketal.（2001）.AnonlineardynamicmodeltoinvestigatethedynamiccharacteristicsofthegrindingprocesswasproposedbyZhangetal.（2005）.AchipthicknessmodelbyGopalandRao（2004）wasdevelopedforassessmentofsiliconcarbidegrindingbyincorporatingthemodulusofelasticityofthegrindingwheel.Basedonthestatisticaldistributionofundeformedchipthickness,predictivemodelsforgrindingforceandpowerwereproposedbyHeckeretal.（2007）.Tocontrolthedepthofcut,agrindingforcebasedmodelmaybeused（HekmanandLiang,1999）.Arealtimegrindingforcemodel（JenkinsandKurfess,1999）wasusedtocontrolthegrindingnormalforce.Theslidingforcemodelforsurfacegrinding（Tangetal.,2008）involvedprocessparametereffectsonfrictioncoefficient.Forpowerrequirementandsurfacefinishprediction,aGA-fuzzymodelmaybeused（Nandiaetal.,2004）.

1.2.Formerror

Itcanbeseenthatthemodelspreviouslyexaminedaremainlyrelatedtoroughness,power,heat,chipthickness,anddynamiccharacteristicsinagrindingprocess.Formerror,asoneofthemostcriticalqualityelementsofaworkpieceamongsize,formandroughness（Whitehouse,2002）,remainsadifficultissue.Veryfewexistingstudiesonformerrorpredictionforsurfacegrindingcanbefound.

1.3.Partialremovalproblem

Foratypicalsurfacegrindingprocess,duetoinitialsurfaceprofileerrorandmisalignmentinvolvedinworkpieceandworkpiecemounting,machiningusuallybeginswithpartialremovalofworkpiecesurface（Fig.1）.Ifthisconditionisnotexaminedasinourpreviousmodel（HuangandGao,2010）,wewillexperienceasignificantamountoferrorinpredictionofformaccuracyinthemachiningprocess（HuangandGao,2010）.Assuch,ourabilityforprecisioncontrol（GaoandJones,1993）,whichisverydesirable（Ludwicketal.,1999;KimandTrumper,1998）,couldbereduced.

1.4.In-processsensingproblem

Inourpreviousmodel（HuangandGao,2010）,in-processmeasurementdevelopedbyGaoetal.（2009,2010）wasnotavailableandtheinitialworkpieceprofiley0（x,z）wasassumedtobezero.Becauseofthis,wecouldnotaccuratelydeterminesystemparametersandwecouldnotacquireaccuratesurfaceforminformation,whichcouldpreventusfrommodelingthemachiningprocessforprecisioncontrol（GaoandJones,1993;HuangandGao,2010）,whichismuchdesired（Ludwicketal.,1999;KimandTrumper,1998）.Inourprevioustest（HuangandGao,2010）,themachiningprocesswasnotfullymodeled.Themaximummodelingerrorwasupto19.9%（HuangandGao,2010）.

1.5.Improveddiscretesystemmodelandin-processsensingtechnique

Inordertosolvetheaboveproblemstoachievehighermodelingaccuracyforprecisioncontrol（GaoandJones,1993;HuangandGao,2010）,anewin-processsurfaceformprofilemeasurement

prototypesystem（Gaoetal.,2009,2010）wasdeveloped.Aftertheuseofthein-processformerrorsensor,accurateparameterestimationandtheconditionofpartialremovalwillbeincludedinan

improveddiscretesystemmodel,whichisproposedforthesurfacegrindingprocess.Theimproveddiscretesystemmodelwillbeexamined.Theproposednewmodelandin-processsensingtechniquewillbeusefulunderbothpartialandfullremovalconditions.Itcanbeusedtopredictworkpieceformerrorthroughoutagrindingprocess.Inaddition,theimprovedmodelwillallowadvancedcontroltoachievehighefficiencyinultraprecisionandnanoprecisionmachining.

2.In-processformerrormeasurement

Foraccurateparameterestimationandidentificationofpartialremovalcondition,in-processmeasurementofformerrorprofileyn（x,z）（Fig.1）willbeusedtoenhancemodelingaccuracy.Withoutthein-processformerrormeasurement（Figs.1and2）,theworkpiecehastoberemovedfromthemachiningpositiontomeasureofflinetoobtaintheinitialworkpieceprofiley0（x,z）.Asignificantamountoferrorwascausedastheworkpiecehastobemountedagainaftertheofflinemeasurement（HuangandGao,2010）.ThisisparticularlytruefornonferrousmaterialssuchasAl,Cu,andSiwhicharecommonlyusedwheremagneticchuckwillnolongerbefunctional.

2.1.Prototype

Toavoidthescratchprobleminthecontactapproach（Whitehouse,2002）,anewopticalin-processsurfaceformmeasurementprototypesystemhasbeendevelopedandhasbeenusedonarealsurfacegrindingmachine（Fig.2）（Gaoetal.,2009,2010）.Themeasurementsystemismainlymadeupofatriangularlasersensor,applicatorandairpiece.ThetypeofthelasersensorisCyberOpticsDRS300.Ithasaresolutionof50nmandisusedtomeasureworkpiecesurfaceformprofileyn（x,z）（Figs.1and2）.Theapplicatorisforreducingtheamountof

coolantnearthemeasurementregion（Fig.2）.Thiswillbebeneficialforestablishingatransparentwindowforopticalmeasurement.Theairpieceistoallowanairbeamtobeejectedandthenimpingedontheworkpiecesurfacetoremovecoolant.Thesystemisthefirstoneofthetypeforworkpieceformprofilemeasurementforwhichthecoolantproblemthatpreventsuseofhighprecisionopticalapproachwassolved（Gaoetal.,2010）.

2.2.Opaquebarrierandvibration

Duetotwokeyproblems,whichareopaquebarrierandvibration（Gao