刀具状态监测系统高速车削振动应变分析外文翻译Word格式文档下载.docx

刀具状态监测系统高速车削振动应变分析外文翻译Word格式文档下载.docx

《刀具状态监测系统高速车削振动应变分析外文翻译Word格式文档下载.docx》由会员分享，可在线阅读，更多相关《刀具状态监测系统高速车削振动应变分析外文翻译Word格式文档下载.docx（12页珍藏版）》请在冰豆网上搜索。

国外作者：

H.Chelladurai，V.K.Jain，N.S.Vyas

文献出处：

《TheInternationalJournalofAdvancedManufacturingTechnology》,2008,37（37）:

471-485

字数统计：

英文2186单词，11662字符；

中文3552汉字

外文文献：

Developmentofacuttingtoolconditionmonitoringsystemforhighspeedturningoperationbyvibrationandstrainanalysis

1Introduction

Inthelastthreedecadesorso,therehavebeentremendousimprovementsandtechnicalrevolutionsinmanufacturingindustries,namelycomputerintegratedmanufacturingprocess,robotcontrolledmachiningprocesses,andothers.Todaycustomerdemandshighqualityproductsforlowestpossibleprice.Tomeetcustomers’suchdemandsandtofaceglobalcompetition,modernindustriesarefacingvariouschallengestowardsachievinghighdimensionalaccuracywithmirrorsurfacefinishontheproducts.Toachievesuchgoalsthemanufacturersarefocusingonthetechnicalproblemsnamely,howtoachieveuninterruptedautomatedmachiningprocessforlongerdurationwithleasthumansupervision.Cuttingtoolwearconditionmonitoringisanimportanttechniquethatcanbeusefulespeciallyinautomatedcuttingprocessesandunmannedfactoriestopreventanydamagetothemachinetoolandworkpiece.Inanymetalcuttingoperation,oneofthemajorhurdlesinrealizingitscompleteautomationisthatofthecuttingtoolstateprediction,wheretool-wearisacriticalfactorinproductivity.Cuttingtoolconditionmonitoringcanhelpinon-linerealizationofthetoolwear,toolbreakage,andworkpiecesurfaceroughness.

Researchersandengineershavebeentryingtoevolveacuttingtoolconditionmonitoringsystemwithhighreliability.Thereisaneedforreliable,universalcuttingtoolconditionmonitoring（TCM）system,whichissuitableforindustrialapplications.Varioussensingtechniqueshavebeenreportedintheliteraturebyvariousinvestigatorswhichdealwiththeissuesofdetectingedgechipping,fracture,toolwearandsurfacefinish.Manysensorswereadoptedintheareaofmetalcuttingtoolconditionmonitoringsystemnamely,touchsensors,powersensors,acousticemissionsensors,vibrationsensors,torquesensors,forcesensors,visionsensorsandsoon.Inanyautomationprocess,sensorsandtheirsignalinterpretationplayanimportantrole.Theprocessingandanalysisofsignalsisimportantbecauseitwillimproveproductioncapacity,reliability,reduceddowntimeandimprovedmachiningquality.Sensorsandtheirutilizationwereimplementedinmanyareaslikemachinetool,automotive,andtoolmanufacturing.Byrneetal.reportedthat46%ofthesensorsmonitoringsystemswerefullyfunctional,16%hadlimitedfunctionality,25%ofthesystemswerenon-functionalduetotechnicallimitationsand13%werereplacedbyorswitchovertoalternatesystems.Itshouldbenotedthat,insensorsbasedsystemsthemostcriticaldecisionispredictionofcuttingtoolconditionusingsignalresponse.Inmanycaseswronginterpretationofthesensorsignalsbyanoperatorleadstothewrongdecisiontoswitchoffthemachinetoolwhichaffectsthequalityoftheproductaswellasproductionrate.Thetrainingofthepersonnelalsoplaysavitalroleinsuccessfulimplementationoftoolconditionmonitoringsystems.

Generally,machiningprocessesarenon-linearandstochasticinnature,anditisdifficulttobuildamathematicalmodel,whichrequiressuitableassumptionsandmaynotbematchingwithrealworldmetalcuttingprocess.AnintensiveresearchhasbeencarriedoutrelatedtoTCMsystemcoveringvariousmetalcuttingprocessessuchasturning,milling,drillingandgrinding,overthepasttwodecadesorso.Agoodcuttingtoolconditionmonitoringsystemshouldbecharacterizedby（a）fastdetectionofimpactorcollisions,（i.e.,unwantedmovementbetweentoolandworkpiece,ortoolandanyothercomponentofthemachinetool）,（b）toolchipping（cuttingedgebreakage）,and（c）gradualtoolwear（craterandflank）causedbyabrasionduetofrictionbetweencuttingtoolandworkpiece（flankwear）andcuttingtoolandchip（craterwear）.

Toolwearsensingtechniquesarebroadlyclassifiedintotwocategories:

directandindirectasshowninTable1.Thedirecttoolwearmonitoringmethodscanbeappliedwhencuttingtoolsarenotincontactwiththeworkpiecelikeradioactive,microscope,cameravisionandsoon.However,directmethodsofmeasuringtoolwearhavenotbeeneasilyadaptableforshopfloorapplication.Theyarenotsuitableforon-lineconditionmonitoringsystemhowevertheycanbeeasilyappliedtooff-linemeasurementsanditconsumesmoretime.Indirecttoolsensingmethodsuserelationshipbetweencuttingconditionsandresponseofmachiningprocesswhichisameasurablequantitythroughsensorsignalsoutput（suchasforce,acousticemission,vibration,orcurrent）andmaybeusedtopredicttheconditionofthecuttingtool.Theseindirectmethodsareusedextensivelybyvariousresearchersandthedetailedanalyseshavebeencarriedoutinthepasttwodecades.Theseindirectmethodscanbeimplementedtoanindustrialproblem,buttheyhavealowersensitivitycomparedtodirectmethods.Nowadays,availabilityofcomputationalpowerandreliabilityofelectronicshelpinthedevelopmentofareliableconditionmonitoringsystembyusingindirectmethods.However,aprobleminTCMsystemisselectionofpropersensoranditslocation.Thesensorshavetobeplacedascloseaspossibletothetargetlocation（closetothetooltip）beingmonitored.

ItisinterestingtonotethatanindirectTCMsystemconsistsoffoursteps:

（i）collectionofdataintermsofsignalsfromsensorssuchascuttingforce,vibration,temperature,acousticemissionand/ormotorcurrent,（ii）extractionoffeaturesfromthesignals,（iii）classificationorestimationoftoolwearusingpatternrecognition,fuzzylogic,neuralnetworks,orregressionanalysis,and（iv）developmentofanadaptivesystemtocontrolthemachiningprocessbasedoninformationfromthesensors.

Therehavebeenmanyinvestigationsontoolwearbasedonperiodicmeasurementsofwearlevelsusingopticalmicroscope.Inthepresentstudy,artificialwearhasbeencreated（externally）usingelectricdischargemachining（EDM）processinacontrolledmanner.Thisissimilartoarealflankwearexperiencedbythetoolduringmachiningprocess.Usingthis（artificiallyworn）cuttingtool,cuttingexperimentsareperformedandsignalsarecaptured.Thispaperaimstodevelopasystematicfeatureextractionprocedurefromtheoutputresponses（accelerometersandstraingauges）ofamachiningprocess.Thesefeaturesarethenfedasinputstothebackpropagationfeedforwardneuralnetworks,forclassificationorestimationoftoolwear.Sameexperimentaldatawerefedasinputstotheanalysisofvariance（ANOVA）totestthelevelofsignificance.Theregressionempiricalmodelshavebeendevelopedandvalidatedwithexperimentalresults.

2Creationofartificialwear

Thecuttingtoolfailurecanbeclassifiedintotwocategories:

（a）gradualwearduetolossofmaterialonanasperityormicrocontactbetweencuttingtoolandworkpiecewhichcausesfrictionandleadstotoolfailureafteracertaintimeinterval,and（b）prematurecuttingedgefailureduetochippingwhichoccurssuddenlyduetoimproperselectionofmachiningparametersortoolmaterialdefect,orboth.Ifacuttingtoolapproachestowardstheendofitsusefullife,surfacequalityofthemachinedworkpieceislikelytodeteriorate.Characteristicsofthesurfacetopographyofamachinedworkpiecedependontheconditionofthecuttingtoolincludingtoolgeometry,cuttingtoolmaterial,workpiecematerial,cuttingconditions（withorwithoutcooling）andmachiningparameters（cuttingspeed,feedrateanddepthofcut）.Ingeneral,thesinglepointturningtoolissubjectedtodifferenttypesofwearsuchasflankwear,craterwear,nosewearandchipping.Outofthesewears,flankwearisconsideredinthepresentworkalthoughothersareequallyimportantunderdifferentmachiningconditions.Whilecreatingartificialflankwear,thefollowingimportantcuttingtoolgeometricalparametersaretakenintoconsiderationlikerakeangle,clearanceangle,lengthoftheflankwearandradialwearlength.

Fig.1Flankwearforms

Therelationbetweenflankwearandradialwearisgiven（Fig.1）by

（1）

where,

isradialwear,

isflankwear,

isclearanceangleand

israkeangle.

AccordingtoEq.

（1）,thevalueofdepthofcut（

）andradialdistance（

）arecalculatedtocreateflankwearintherangeof0.2mmto0.5mm.EDMexperimentalsetupisusedtocreateartificialflankwear.ThereplicaoftheflankwearisfirstproducedinthecopperrodbyturningoperationandthenthecopperrodisusedintheEDMprocessasatool（cathode）toreplicatetheflankwearonthecuttingtool.DuringEDM,thetungstencarbidecuttingtool（usedasaworkpieceinEDMoperation）ismadeasananode.MachiningparametersvaluesarechosenintermsoffeedrateaswellasdepthofcuttocreateexactshapeofflankwearasshowninFig.2.Themeasurementofwearcanbecarriedoutusingoneofthethreecategoriesofsensors,namelyproximitysensors,radioactivesensorsandvisionsensors.Mostoftheresearchersuseanopticalmicroscopetodeterminewornoutareas,likeflankwearlengthandcraterwearlength.

Inthisstudy,opticalUSBportmicroscope（Scalar）isusedtocapturetheimageofwornoutareawithmagnificationfactorof50x.Beforemeasurements,insertismountedonastand,madeofPerspexhavingincludedangleof55°

asshowninFig.3.Itc