预加工数控轮廓确认英文.docx
《预加工数控轮廓确认英文.docx》由会员分享,可在线阅读,更多相关《预加工数控轮廓确认英文.docx(25页珍藏版)》请在冰点文库上搜索。
预加工数控轮廓确认英文
PRE-MACHININGCNCCONTOURVALIDATION
TonySchmitz,JohnZiegert
UniversityofFlorida,MachineToolResearchCenter
Gainesville,FL32611
Abstract
ThecurrentprocedurefollowedtomanufactureanewpartbyCNCmachiningistowritethepartprogram,
machineatestpartandmeasurethetestpartforconformancetotherequireddimensionsandtolerances.Ifthetest
partdimensionsarenotcorrect,thepartprogramismodifiedandtheprocessrepeateduntilasuccessfulpartis
machined.Inmanyapplications,suchastheaerospaceindustry,wherematerialcostandmachiningtimearehigh,
thisiterativeprocessbecomeseconomicallyunacceptable.
ResearchhasbeenconductedtotestthefeasibilityofusingtheLaserBallBar(LBB),aspatialcoordinate
measuringdevice,tomeasuredynamiccontinuous-pathcontoursofCNCpartprogramstomicrometeraccuracy
priortomachining.Inthisway,avirtualtestpartcanbemeasuredandcomparedtothedesigndrawingsto
validatetheCNCpartprogram.Thisreducesoreliminatesthecostlyandtime-consumingstepsinvolvedinthe
machiningofphysicaltestparts.
ThispaperoutlinesthetestingmethodandresultsacquiredusingoneLBBtomeasuredynamicpartpaths
employingsequentialtrilateration.Acircularcontourwasmeasuredusinganencodertriggerfordatacapture.
Theradialerrormotionsofthespindleusedtogeneratethecircularcontourwerealsomeasuredusinga
capacitanceprobenesttoverifytheLBBresults.ComparableerrorwaveformsbetweentheLBBandcapprobe
measurementsverifiedthepossibilityofusingtheLBBtomeasuredynamiccontinuous-pathcontours.Futurework
usingthreeLBBssimultaneouslyisalsooutlined.
Keywords:
LaserBallBar;CNCmachining;dynamicmeasurements
Introduction
OneofthemostimportantusesofCNCmachinetoolsisthecuttingofcomplex
continuouspartpaths,orcontours.Incontinuous-pathnumericalcontrolsystems,thereis
contactbetweenthecuttingtoolandworkpiecethroughoutthepartpathwhileuptofiveaxesare
inmotion.Therefore,thefinalworkpiecedimensionsaredirectlyrelatedtothepositional
2
relationshipbetweenthetoolandworkpiece.Theabilitytomonitorthisrelationshipand
predictthefinalpartdimensionsisimportantfortoday’smanufacturingengineer.
ForanewpartorproductionruntobemanufacturedwiththeuseofCNCmachinetools,
thecurrentprocedureistowritetheCNCpartprogramusingtheengineeringdrawings,execute
thepartprogramtomachineatestpartorprototype,andtheninspectthetestpart,normallywith
theaidofacoordinatemeasuringmachine,tocheckforconformancetodesigntolerances.This
feedbackoftheactualpartdimensions(withanadequatedegreeofprecision)iscurrentlythe
onlywaytocertifytheperformanceoftheCNCprogram.Ifthetestpartdoesnotmeetthe
specifiedtolerances,asisoftenthecaseforafirsttrial,theCNCprogramismodified,another
testpartismachinedandtheprocessrepeated.
Thisiterativeprocessmaybeacceptable,althoughnotefficient,insituationswherethe
materialisinexpensiveandmachiningtimeisshort.However,inmanycases,suchasthe
aerospaceindustry,thematerialiscostly(expensiveforgingstobemachined)andmachining
timeisdramaticallyincreased.Insuchinstances,thisiterativeprocessofproducing
dimensionallycorrectpartsbecomeseconomicallyunacceptable.
Bothtimeandmoneycouldbesavediftherewereawaytodirectlymeasurethemachine
tool’sdynamiccontouringaccuracyoveranarbitrary3-Dpathwithoutthenecessityof
machininganexpensivetestpart.Atthesametime,theefficiencyoftheCNCmachining
processcouldbegreatlyincreased.TheuseoftheLaserBallBar(LBB),aspatialcoordinate
measuringdevice,totakedynamicmeasurementsofapathcouldreduceorreplacetheneedto
machineandinspectatestpart.Ineffect,thespatialcoordinatesofthedynamictoolpath
measuredbytheLBBcouldfunctionasavirtualtestpart1.
Thisparticularapproachmustbedifferentiatedfromtheconventionalapproachto
3
machinetoolmetrology.Atthistime,thenormalprocedureistomeasurethemachinetoolerror
motionsincludinggeometric,thermalandperhapsprocesserrors.Theseerrorsmaythenbeused
aspre-machiningcompensationinthemachinetoolcontrollerinanefforttogivetheoretically
perfectmotions.Themachine’saccuracyand,therefore,thepartdimensionsarebasedlargely
onthesuccessofthistime-consumingprocess.Althoughthisresearchiscertainlynota
replacementforthisbodyofwork,itisalsonotthesame.Thepurposeofthisworkistotryand
predictthefinalpartdimensionsfortheexecutionofaspecificCNCpartprogramonagiven
machinetoolpriortocuttingthepart,nottomeasurethegeometricorservoerrorsofthe
machinetool.
Background
OneofthemaincategoriesinthemodernevaluationofCNCmachiningcentersisthe
assessmentofthecontouringcapabilitiesofthemachinetool.Theevaluationofthiscontouring
accuracycanbedividedintotwomainclassifications:
post-processandin-processtesting.
Post-processtestingincludesthosetestswhichareperformedaftermachininghasbeen
completed.Themostpopularpost-processinspectiontoolisthecoordinatemeasuringmachine
(CMM).Otherpost-processmethodsusedtoevaluateaCNCmachinetool’scontouring
accuracyinvolvetheuseofeithermasterpartsorwell-definedcontours.Oneexampleofthis
techniqueistheuseofstandardpartpaths,suchasthepartprogramcorrespondingtothe
NationalAerospaceStandardtestpart979(NAS979),tomachineamasterpart.Thispartcan
thenbemeasuredtoevaluateflatness,squareness,parallelism,roundness,etc.Asimilar,but
somewhatmoreefficientmethodhasbeentermedmasterparttracing.Thisproceduresimulates
machiningbyreplacingthetoolwithagageandtheworkpiecewithamasterpartofknown
4
accuracy.Themachineisthenprogrammedtofollowtheidealpathgivenbythemasterpart.
Deviationsfromtheidealpathregisteredbythetoolgagerepresentthecontouringerror2,3.
Theothermaintypeofcontourmeasurementisin-processtesting.Thiscategorycan
thenbesubdividedintoin-processandin-cyclegaging.In-processgagingreferstotesting
carriedoutduringtheactualmachiningprocess,whilein-cyclegagingdefinesmeasurements
takenafterthepartisfinished,butbeforeitisremovedfromthemachine.
In-processgagingisgenerallyachievedbyaddinganalogtransducerstothemachinetool
todirectlyassessthetestpart’ssizeduringthecuttingoperation.Anexampleofin-process
measuremententitledWorkpiece-ReferredFormAccuracyControl,orWORFAC,hasbeen
suggestedbyUdaetal.4Itconsistsofahighlysensitiveopticalsurfacesensorandamicrotool
servotoadjustforchangesinthedistancebetweenthetoolholderandworkpiece.Although
accuracyimprovementswereshownwiththeuseofthisfeedbacksystem,itisstilllimitedto
simplegeometries(cylindricalturningoperations)andrequirescomplextoolingandset-up.
In-cyclegagingcanbeaccomplishedbyreplacingoneormoreofthetoolsinaCNC
machiningcenter’sturretwithmeasuringprobes.Theseprobescanthenbeindexedintothetool
positiontoprobekeyfeaturesofthetestpartwhileitremainsinthemachineand,unfortunately,
isstillinathermallyunstablestate.
AnadaptiveerrorcorrectionmethodhasbeenproposedbyMouetal.whichcombines
bothin-cycleandpost-processtestingwithageometric-thermalerrormodel5,6.Thistechnique
proposestheuseofin-cycleandpost-processgagingtointeractivelydeterminethechangesin
thegeometric-thermalmodelovertime.
AlthoughtheseprocessesprovideworthwhilemeansofevaluatingaCNCmachinetool’s
5
contouringaccuracy,theyallsufferfromthenecessityofmachiningandinspectingatestpart
forconformancetorequiredtolerances.Thepurposeofthisresearchwastoinvestigatethe
possibilityofusingtheLaserBallBarasameasurementtooltodeterminethisCNCpath
accuracywithouttherequirementofproducingacostlytestpart.Commercialproductswhich
aremovingtowardthisgoalincludetheHeidenhain2-DgridplateandRenishawBallBar.
Thesetools,however,arepathlimited.TheBallBarallowsonlycircularorhemisphericalpaths
andrecordsjustradialdeviation.Thegridplatecanmeasureonlyplanarpartpaths.For3or5-
axispartpaths,theLaserBallBarsystemisrequiredtodynamicallymeasurethesecontours.
LaserBallBar
TheLBBisaprecisionlineardisplacementmeasuringdevicedevelopedbyZiegertand
MizeattheUniversityofFlorida7.Itconsistsofatwo-stagetelescopingtubewithaprecision
spheremountedateachend.Aheterodynedisplacementmeasuringinterferometerisaligned
insidethetelescopingtubeandmeasurestherelativedisplacementbetweenthetwospheres.See
Figure1.TheLBBhasbeenshowntobeaccuratetosub-micrometerlevelsduringstatic
measurementsofspatialcoordinates8.
Onceinitialized,theLBBusestrilaterationtomeasurethespatialcoordinatesofpoints
alongaCNCpartpath.Thesixsidesofthetetrahedronformedbythreebasesockets(attached
tothemachinetable)andatoolsocket(mountedinthespindle)aremeasuredand,bygeometry,
thecoordinatesofthetoolpositioncanbecalculated.
Insequentialtrilateration,thesamepartpathistraversedthreetimes,measuringthe
lengthsofoneofthebase-to-toolsocketlegsatafinitenumberofpointsduringeachrepetition.
SeeFigure2.Notethatthetoolsocketmustbeinexactlythesameposition(foragivenpoint)
6
foreachofthethreemeasurements.Ifthetoolsocketwere