Microsoft Word 文档.docx

Microsoft Word 文档.docx

《Microsoft Word 文档.docx》由会员分享，可在线阅读，更多相关《Microsoft Word 文档.docx（11页珍藏版）》请在冰豆网上搜索。

MicrosoftWord文档

FailureofRapidPrototypeMoldsduringInjectionMolding

J.S.Colton’,J.Crawford’,G.Pham’,V.Rodet’

CenterforPolymerProcessing,SchoolofMechanicalEngineering,

GeorgiaInstituteofTechnology,Atlanta,Georgia,USA

SubmittedbyK.K.Wang（I）,Ithaca,NewYork,USA

1

Abstract

Stereolithographyrapidtoolingmaterials（typicallyepoxy-basedphotopolymers）showgreatpromise

forinjectionmoldingoflimitednumbersofparts,greatlyreducingthetimetoproduct.Yet,theypresentchallengestodesignersbecauseoftheirstrength,thermalcharacteristics,andshorterlifetimesascomparedtoothermoldmaterials.Thispaperpresentsmodelsoftheforcesgeneratedduringtheinjectionmoldingcyclesoastoevaluatethesuitabilityofrapidtools.Themodelscomprisethermalandmechanicalloadingduringinjection,coolingshrinkage,andejection.Theeffectsofprocessconditionsonmaterialpropertiesarestudied.Experimentalresultsarepresented.

Keywords:

rapidprototyping,injectionmold,failure

1INTRODUCTION

RapidToolingshowsgreatpromisefortheinjectionmoldingoflimitednumbersofparts,greatlyreducingthetimetoproduct,e.g.,[I-61.However,stereolithography（SL）rapidtoolingmaterials（typicallyepoxy-basedphotopolymers）presentchallengestodesignersbecausetheirstrengthandthermalcharacteristicsarenotasgoodastraditionalsteelmolds.Inaddition,longercycletimesarerequired.InsomecasesoutsidetheUSA,fillersareaddedtoincreasestrength,butmayinterferewithcure.SLinjectionmoldshavemuchshorterlifetimesthanothermoldmaterials,onereasonbeingthattheyareoperatedattemperaturesclosetotheirglasstransitions（Tg）.Also,thestereolithographyprocessoflayeredmanufacturingandthedraftanglerequirementsgeneratestairsteppingonthesurfaceofthepart,whichincreaseejectionforcesandcanproducestressconcentrations.Thispaperpresentsmodelsoftheforcesgeneratedduringthemoldingcyclesoastoevaluatethesuitabilityofrapidtoolsforinjectionmolding.

Thematerialpropertiesoftherapidprototypingmoldmaterialsareevaluatedwithmechanicaltesting（tensileandflexural）anddynamicmechanicalanalysis（DMA）attemperaturessimilartothosepresentinthemoldduringinjectionmolding.MechanicaltestsshowthattheSLmoldingmaterialfailsinabrittlemanner,withlittleifanyfatiguebehavior.Somechangesinmechanicalpropertiesareobservedduetomechanicalandthermalcyclingofthemoldmaterials,eventhoughtheyarefullycuredduringprocessing,asevidencedbydifferentialscanningcalorimetry（DSC）.Theseprovidetherequiredpropertiesforthecalculations.

Duringinjection,theflowofthepolymerintothemoldcavitycancausemoldfeaturestodeflectandbreak,particularlyafteranumberofcycleswhenthemoldhasheatedup.Theseforces,coupledwiththetemperaturedependentmoldproperties,aremodeledusingfiniteelementtechniquestodeterminetheextentofthemold

featuredeflectionorpossiblemoldfailure.Experimentsvalidatethesetheoreticalresults.

Duringcooling,aparttypicallywillshrinkontothecoreofamold.TheinherentstairstepgeometryoftheSLmoldismodeledsothatitsgeometrycanbedeterminedtheoretically.Bothofthesecontributionsareincorporatedintoanejectionforcemodel.Finiteelementtechniquesareusedtocalculatetheejectionforces.Experimentaldataforejectionforcesarepresented.

2THEORETICALMODELING

2.1InjectionPhase

Thefirstfeaturestudiedinthispaperisafree-standingribthatismodeledasacantileverbeam（Figure1）withalinearlydecreasingdistributedload（i.e.,higherloadatthefreeend）,whichrepresentsloadingbytheincomingpolymerflow.Themaximummoldfeaturedeflection,y-,ispredictedusingafiniteelementmethodandtheelementaryequationforcantileverbeamdeflection（Equation1）.Additionally,thestress,o,experiencedinsidethefeatureispredictedusingafiniteelementanalysisandtheelementaryflexureformula（Equation2）.

wherePisapointloadattheendofthebeam,Lisitslength,Wisadistributedloadonthebeam,Eisthemodulus,Iisthesecondmomentofarea,Misthebendingmoment,andcisthehalfthicknessofthebeam.TheforcesandmomentactingonthecantileverbeamduetotheincomingpolymerflowarepredictedbyC-Mold,aninjectionmoldingsimulationprogram.ThedetailsofthefiniteelementanalysisarepresentedinSection2.3.Themoldfeatureispredictedtofailifthestressexceedsitsmaterial’smaximumflexuralstress.

2.2CoolingandEjectionPhases

Themodelforejectionforceisdevelopedbycombiningtheeffectsofthemajorcontributingfactors:

thethermalshrinkageandtheinherentstair-stepprofile（Figure2）.

Mathematically,

Figure1:

Samplefeatureandcavitygeometry

whereFfric,thermandFdef,dalrdenotetheejectionforcecomponentsduetothermalshrinkageandtostair-stepping,respectively.Forageneralmoldwithacorefeature,themoldedpartcoolsdownandcontractsontothecore,creatingcontactpressureattheinterface.Thecontactpressuregeneratesfrictionduringejection,Ffric,therm,whichmustbeovercometomakeejectionpossible.Foramoldwithacorefeature,Ffric,thermcanbeestimatedusingathick-walledvesselapproximation[I].ThelayerednatureoftheSLprocessleadstothestair-stepphenomenondepictedinFigure2.Thisstair-stepprofilecreatestheundercutoroverlap6thatpreventsthemoldedpartfromejecting.Thestair-stepprofileispreservedthroughmultipleejectioncycles[2].Hence,thepartandmoldmustdeformelasticallytoovercometheoverlap（Fdef,stalr）inordertomakeejectionpossible.

InordertoformulateFdef,stalr,theoverlap6wasquantifiedmathematically.Toachievethisgoal,thestair-stepprofilewasmodeledusingtheequationofthehatchedbulletprofile[3].Thearithmeticaverageroughness（Ra）,andconsequently,theoverlap6ofaSLtoolsurfacecanbeestimatedasfollows:

where0isthedraftangle,CL~isthelinewidthcompensationorbeamcompensation,Iisthebuildlayerthickness,andOCistheover-cure.CL~andOCdepend

onresintype,lasertype,andmachineset-up.BycombiningFfrlc,thermandFdef,stalr,theejectionforceequationbecomesthefollowing:

wherePthermisthecontactpressureduetothermalshrinkage,bqisanequivalentfrictioncoefficient,pisthefrictioncoefficientbetweenmoldandpartmaterials,SAisthecontactareaalongtheejectiondirection,ATisthechangeintemperaturefrominjectiontoejection,risthehydraulicradius=（2*area/perimeter）,EisYoung'smodulus,visPoisson'sratio,andm,paresubscriptsdenotingmoldandpart,respectively.InregardstoATp,theYoung'smodulusofthepart'smaterialisonlysignificantattemperaturesbelowitsglasstemperatureTg,p.Therefore,thethermalstrainduetopartcooling

shouldbeincludedonlyafterthepartcoolsbelowTg,p.Inotherwords,ATp=Tg,p-Te,pwhereTe,pdenotestheejectiontemperatureofthepart.

2.3FiniteElementModelingandAnalysis

FEanalyseswereperformedtovalidatetheforceequations（Equations2and5）andtheexperimentalresults,andforuseasindependentmodelstopredictinjectionandejectionforcesforthecaseswheretheforceequationsmightbeinsufficient,e.g.,moldsoflargersizeormorecomplexgeometry.Acoupled-fieldsequentialtechniquewasemployedtoperformthermalanalysisandthesubsequentstructuralanalysisforthemold/partset.Forinjection,C-Moldwasusedtodeterminethethermalandpressurestates.ANSYSthenwasusedtodeterminethestressstate.Forejection,3-Dmodelingofthemold/partsetwasdoneusingProEandANSYS.ThermalanalysisthenwasperformedinANSYStoobtainthetransienttemperaturedistributionwithinthemold/partset.Thethermalresultswereusedasinputtothestructuralanalysistodeterminethestressstateatejection.Thecontactstressthenwasobtained.

3EXPERIMENTALPROCEDURES

3.1MaterialCharacterization

Experimentswereperformedtodeterminetheeffectofprocessconditionsonthemechanicalpropertiesoftwostereolithographyresins:

CibaTool7510andDSMSomos7110.DynamicMechanicalAnalysis（DMA）（singlecantilever35x12~2mm）,compactdogbone（ASTMD638-94b:

typeII）,andfracturetoughnessbend（ASTME399-90;thickness（B）of12mm）specimenswerebuiltinthreeorientations:

top,bottom,andside.Thesedesignationsrefertothelocationofthelayering（Figure2）onthepartresultingfromthebuildprocess.AfterbuildandonehourinanUVchamber,partswerepostcured

fortwohoursinathermalovenat8OoC.Samplesweretestedinanotfurtheragedcondition（NA）;athermallyagedcondition（TA）-subjectedtosixmorehoursat8OoC;andamechanicallyagedcondition（MA）-subjectedto1000bendingcyclesof1mmamplitudeinaDMAat3OoC.Thelatterconditionssimulated,inamanner,theeffectsofprocessing.Amodel2890DMAfromTAInstrumentsandamodel4446universaltestingmachine（UTM）withtemperaturechambermodelA74fromlnstronwereusedforthetests.DMAtestswereperformedwitha3'C/minsweeprate,20pmdisplacement,and1HzfrequencytodetermineT,,E'（storagemodulus）,andE"（lossmodulus）.FatiguetestingalsowasperformedontheDMA.Itwasperformedfor300bendingcyclesof1mmamplitudeforthe7510resin,anduntilfailureforthe7110resin.TensiletestingwasperformedontheUTMat1.5mm/minwitha25mmgaugelengthandanextensometer.

3.2InjectionMoldingExperiments

Parts（Figures1and3）wereinjectionmoldedonaSumitomoinjectionmoldingmachine（SG75-CI60-Mlll）usingChevronMC3600polystyrene.Themoldswerebui