flow visualization and simulation of the filling process ... › ws › files › 140502737 ›...

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

You may not further distribute the material or use it for any profit-making activity or commercial gain

You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: Jul 04, 2020

Flow visualization and simulation of the filling process during injection molding

Guerrier, Patrick; Tosello, Guido; Hattel, Jesper Henri

Published in:C I R P - Journal of Manufacturing Science and Technology

Link to article, DOI:10.1016/j.cirpj.2016.08.002

Publication date:2017

Document VersionPeer reviewed version

Link back to DTU Orbit

Citation (APA):Guerrier, P., Tosello, G., & Hattel, J. H. (2017). Flow visualization and simulation of the filling process duringinjection molding. C I R P - Journal of Manufacturing Science and Technology, 16, 12–20.https://doi.org/10.1016/j.cirpj.2016.08.002

1

FLOWVISUALIZATIONANDSIMULATIONOFTHEFILLINGPROCESSDURINGINJECTIONMOLDING

P.Guerriera,*,G.Toselloa,J.H.Hattela

aDepartmentofMechanicalEngineering,TechnicalUniversityofDenmark,ProduktionstorvetBuilding425,2800Kgs.Lyngby

*Correspondingauthor,emailaddress:pagu@mek.dtu.dk

AbstractTodirectlycompareexperimentalmoldingsfromaninjectionmoldingmachinewithsimulations,aspecialmoldhasbeenproducedwithaglasswindow.The injectionplane isperpendicular to theopeningandclosingplanes,inorderforthe55mmthickglasswindowtobeeasilyvisiblefromtheside.Ahighspeedcamerarecording500framespersecondwasemployed,andthemoldhadthreethermocouplesandtwopressuresensorsinstalled.Themoldedpartisa2mmthickplatewitha0.5mmthinsection,whichcreatesacharacteristicV‐shaped flowpattern.Twodifferentmaterialswereemployed,namelyABSandahighviscosityPC.Simulationswereperformedusingtheactualmachinedataasinput,includingtheinjectionscrewacceleration.Furthermore,thenozzleandbarrelgeometrieswereincludedasahotrunnertocapturetheeffectofcompressibilityofthematerialinfrontofthescrew.Thesetwohadsignificanteffectsonthefilling times and injection pressure calculated by the simulations. Other effects investigated includedtransientthermalmanagementofthemold,pressuredependentviscosityandwallslip,buttheireffectwerenotremarkably large in thiswork.Theobtainedsimulationresults showeddeviationswithin10‐30ms(relativedeviationintheorderof5‐10%)fortheABSandslightlymoreforthehighviscosityPCintherangeof100‐500ms(relativedeviationintheorderof20‐30%)ontimingsbetweendifferentsectionsduringfilling.

Keywords:GlassMold,HighSpeedCamera,InjectionMolding,Simulations

1 IntroductionDirectflowvisualizationwiththeuseofahighspeedcameraisaknownmethodologytoanalyzetheflowbehavior in injectionmoldingandotherprocesses. Inseveralstudiesexperimentshavebeenconductedwithdifferentkindsofcamerasetupsinthemoldingtools.Yangetal.[1]haveemployedflowvisualizationusingaquartzglasswindowinamicroinjectionmoldingtooltoinvestigatetheflowofmicropartswithhighinjectionspeed,usinghighspeedcameraabletocapturevideoswith1000framespersecond.Themoldcontainedareflectionmirrortoallowspacefortheplacementofcameraandthermocouples.Themainpurposeoftheinvestigationwasthemicroinjectionmoldingprocesscharacterization.Otherstudieswhichalso use a reflectivemirror typemold construction are:Nian et al. [2] use previouslymentionedmoldhoweverthistimecombiningtransparentandcoloredmaterialtogetfurtherinformationoftheflow.Hanetal.[3]useahighspeedcameracapableofrecordingat13,500framespersecondandamicroscope,toinvestigatethefillingofmicrogroovesonaglassprisminsert.Hasegawaetal.[4]useaglassinsertwith

2

branchingrunnersinordertoinvestigatetheinfluencesof inertiaforcesonmeltfillingbehaviorathighinjectionspeed.Layseretal.[5]investigatehesitationandemphasizestheuseofthepackingphaseduetotheuseofawindowmadeoffusedquartzglass,whichwasabletowithstandapackingpressureofupto50MPa. Spares et al. [6] havemodified a thermal imaging camera to increase the frames per seconds, toinvestigatethethermalfieldduringinjectionofmicromolding.Whitesideetal.[7]havealsoinvestigatedmicromoldingwithahighspeedcameraandsapphireglassinsertinordertostudydifferenteffectsinmicroinjectionmolding.Yokoietal.[8]useahighspeedcamerawithamoldwithabuildinmicroscopeinordertoinvestigatemicrocavitiesandtheeffectofinjectionspeed.Allthementionedworksofar,usesareflectivemirrorandnosimulationcomparison.

Otherresearchesinthefieldofflowvisualizationininjectionmoldinghavebeenperformedusingadifferentmolddesignthat,insteadofusingareflectivemirror,includedatransparentinsertdirectlyappliedtothemoldcavitywhichisorientedparalleltotheinjectiondirection,sotheflowcanbedirectlyfilmedfromthesideoftheinjectionmoldingmachine.Yanevetal.[9][10]appliedasapphireinserttothistypeofmolddesign.Theresearchershavealsoperformedasimulationcomparisontoverifythefillingperformanceofthesoftware.Itwasstatedthatthesimulationiscapableofgivinggoodoverallagreementwiththeobtainedflowpatterns,butsignificantdifferencewasfoundduringthebeginningofthefillingphase.Furthermore,Dvoraketal.[11]alsousedadirectobservationoftheflowwithahighspeedcamerawithoutmakinguseofareflectivemirror,forthecharacterizationofceramicinjectionmolding.

Themainfocusofthisworkwillbeonmakinganexperimentalsetupsimilarto[9][10]i.e.usingaturnedinsertfordirectvisualizationoftheflowinthecavity,whichisorientedparalleltotheinjectionscrewofthe molding machine. However in the present work a more comprehensive study of the influencingparameters on the simulation results is carried out. Using pressure sensors and thermocouples incombinationwithahighspeedcamera, thegoalsare tocaptureandcharacterize the fillingof themoldcavity, aswell as to analyze and compare the flowpatternwith simulation on a timebasis. Deviationsbetweenexperimentsandsimulationpredictionswillbestudied.Differentfeaturesinthesoftwarewillbeutilizedtominimizetheobserveddeviationsinordertoobtainthecorrecttimingsofthefillingpattern.Theuse of a high speed camerawith a glassmold will provide a direct comparison of the filling, with anincreasedaccuracywithrespecttomoreconventionalvalidationtechniquessuchastheshortshotsmethod.

2 Experimental

2.1 GlassmoldIn order to compare simulations directly with the filling of the cavity, a special glass mold wasmanufactured.Theinjectionplaneisalmostperpendiculartotheopeningandclosingplane,inorderfortheglasswindowtobeeasilyvisiblefromtheside.Thereasonfornotbeingturned90°,butinstead82°C,wasduetotheopeningandclosingofthemold,topreventatoocloseandtightfitandpossiblebreakageoftheglass,eitherinopeningandclosingofthemoldorduringthepackingphasebyapplyingtheholdingforce.Thismolddesignmakesitparticularlyeffortlesstoinstallanykindofrecordingdevicetothemold.Theglasswindowismadeofa55mmthickBorosilicateglass(widthandheight60x140mm),anditiscapableofwithstandingatleast130MPainmachinepressureduringinjection,and50MPaduringpacking.Itwas

3

testedtogoupto220MPawhichdidnotbreaktheglass,eventhoughthispressurelevelwasnotusedduring the experiments. Due to the (almost) perpendicular injection plane, and a fairly large air ventthickness(0.02mm),moldingscouldresultinsomeflashduringsomeoftheexperiments.Theexperimentalsettingincludingthetwomoldhalves,theglasswindowandtheplacementofthehighspeedcamera,aswellasaschematicoverviewofthemoldcanbeseeninFig.1.

Fig.1–Twomoldhalvesoftheglassmoldandschematicoverview.

2.2 EquipmentandmaterialsTheemployedhighspeedcameraisanX‐PRIproducedbyAOSTechnologiesAG(Fislisbach,Switzerland)anditiscapableofrecording1280x1024pixelsimagesat500framespersecond.Thecamerawaseasilyinstalledinsidethesafetydooroftheinjectionmoldingmachine,andattachedtothetrackoftheslidingdoor.Someadditionallightingwasalsoinstalledtoimprovethebrightnessoftherecordedvideos.

Themoldisequippedwithatotaloffivesensors.Threeofthesensorsarethermocouples,whichareusedtotimewhenthemeltisatthespecificlocation.Furthermore,twopressuresensorsareinstalled,oneintherunnerandoneinthecavityjustafterthegatelocation.Thetwopressuresensorsareusedtodeterminewhenthepolymermeltpassesthem,andtodirectlycomparewiththesimulatedpressure,inconjunctionwiththehighspeedrecordingsystem.Fig.2showsthelocationofthedifferentsensors.Inadditiontotheinstalledsensors,themachinedataisalsocollectedtobeusedasinputforthesimulationsetup,inordertosimulatetheactualconditionsinwhichtheexperimentsareconducted.

TheinjectionmachineusedisanEngelVictory60witha35mmdiameterscrew.Twodifferentpolymermaterialswere employed for the experiments, the first being a natural colored acrylonitrile butadienestyrene(ABS)gradewithameltflowindex(MFI)of7g/10minandthesecondagraycoloredhighviscositypolycarbonate (HVPC) grade with a MFI of 34.3 g/10min. Viscosity and pvT of both polymers werecharacterizedinordertohaveaccurateinputforthesimulationsalsointermsofmaterialproperties.

4

2.3 TestpartThegeometryusedintheexperimentscanbeseeninFig.2.Itconsistsofan18mmwideand45mmlongrectanglewithathicknessof2mm.Ithasathinsectioninthemiddleof0.5mm,whichproducesaflowwhich will slow down over the middle section, as the frozen layer starts to develop. This creates acharacteristicV‐shapedfrozenfrontoverthethinsection,asthesidesflowfasterduetobeingthicker,hencenotsolidifyingthroughthethicknessandallowingmelttopass.

Fig.2–Thetestpartwithanoverallthicknessof2mm,andacentersectionof0.5mm,notedwithsensorlocation.P1andP2indicatethepressuresensors;T1,T2andT3indicatethethermocouples.

2.4 SetupTheinjectionflowratewasspecifiedinorderforthefilling(i.e.injection)timetobecloseto1s.Thistargetwasachievedbyadjustingtheinjectionramspeed,whichwasclosetoconstantat10mm/sforABSand12mm/sforHVPC.Theaccelerationdistancewas1.54mmforABSand1.50mmforHVPC,resultinginanacceleration timeofapproximately0.15s forbothpolymers.The fillingwas followedby4 sofpackingphase,withaholdingpressureof40MPafortheABSandof50MPafortheHVPC(steppeddownto35MPa).Thecoolingtimewassetat15secondsforbothpolymers.ThemelttemperatureoftheABSwassetto237°Cwithcoolingwaterat17°C,whereasthemelttemperatureofHVPCwassetto315°Cand95°Ccoolingwater.Thecoolingchannelislocatedintheoppositesideofthemoldhalfwiththeglass,togetherwiththeejectorsystem.ThepressuresensorsandthermocouplemeasurementsfromtheexperimentscanbeseeninFig.3.

5

Fig.3–Thethreethermocouplesandtwopressuresensorssignalsduringfillingandpartofthepackingphase.Left:ABS.Right:HVPC.

3 Simulations

3.1 TheoreticalbackgroundFor the flow simulations the principle of conservation ofmass,momentum and energy is used for thegoverningequations:

∙ 0 (1)

∙ (2)

(3)

∙ Φ (4)

Φ Δ (5)

where isthedensity, isthevelocitycomponent, isthepressure, isthetotalstresstensor, istheshearrate, isthebodyforce, isthespecificheat, isthethermalconductivity,Δ istheheatreleased

undersolidificationand istheviscositytakenasthegeneralizedWLF‐Crossviscositymodel:

, ,,

1 ∗⁄ (6)

where isapower‐lawconstantand ∗isatransitionstress. isazero‐shearrateviscosityrepresentedbyaWLF‐typemodeltotakeintoaccountthetemperatureandpressureeffects:

6

,∗

∗ (7)

where , ,and arematerialconstants,and ∗istakenastheglasstransitiontemperaturewhichcanvarywithpressurethrough :

∗ (8)

Fortrackingthefluidflowfrontalevelsetmethodisemployed:

∙ 0 (9)

Where isthevolumefraction,where 0istheairphase,and 1isthepolymermeltphase.Themeltfrontistrackedbytheelementswith0 1.Theequationsaresolvednumericallywithafinitevolumescheme using a 3D unstructured boundary layermesh (see Fig. 4). For the purpose of this study, thecommerciallyavailablesoftwareMoldex3DSolidR13wasemployed.

Fig.4–Top:Hybridmeshconsistingoftetrahedronelementsinthemainbodyandprismelementsattheboundary,tobettercapturelargergradients.Thepurpleelementsareinsidethecavity,andthegreenare

onthesurface.Bottom:Modelofthecavity(green),runner(blue),hotrunner(orange).

3.2 Velocity‐topressure‐controlswitch‐overThehighspeedmoviescangiveaccurateinformationregardingthemeltflowfrontpositionwithrespecttotime.However,onechallengeisrepresentedbythedeterminationofthepoint(intermsofbothflowfrontpositionandtime)atwhichthevelocity‐controltopressure‐control(V/P)switch‐overhappens,sinceitisnotclearlyevidentfromthehighspeedcameramovies.Ontheotherhand,thedeterminationoftheV/Pisnecessary in order to obtain an accurate implementation of the simulation and therefore to be able tocomparethesimulatedflowfrontpatternwiththehighspeedcameravideobothduringthefillingandthepackingphases. Inorder toovercome this issue, theV/P switch‐overpointwas foundbyperforming asimulationwithhighpercentfilledpoint(e.g.98%),andthennoticingthetimeatwhichithitsthefirstandsecondthermocouplesT1andT2.BylookingattheexperimentaltemperaturemeasurementsatT1andT2,andinparticularatthepointsatwhichthethermalsignalrisesfromboththermocouples,thetimeneededby themelt to travel between those two points was found. The two timings t(T1) and t(T2) are then

7

subtracted and the time interval t(T1T2) between the two thermocouples is determined. The timeneededbythemelttotravelfromthethermocoupleT2untiltheV/Pswitch‐overpointisthenfoundbylooking at the pressure sensor P1 signal. P1 and T2 share the exact same position and their signal issynchronizedintime,i.e.t(P1)=t(T2).Thetimeintervalt(P1V/P)betweenP1andV/PisthendefinedasthetimebetweentheP1pressuresignalrise(locatedatsamepointsasT2)untilthereisadropintheP1pressuresignal(i.e.thepointwhenthepackingphaseisinitiated).ThispointischaracterizedbythetimeatwhichtheV/P,i.e.t(T1T2)+t(P1V/P)ishappeninginthesimulation.Thepressureatthattimeisthennoteddown,andusedasthepressureatwhichtheV/Pswitch‐overistakingplace.

ThiswasnotpossibletobeobtainedinthereferencesimulationssimplyduetothefactthatthefillingtimewassofastthattheV/Pswitch‐overwouldnotbetowardstheendofthefillingphase,butinsteadratherlater into thepackingphase.When theaccelerationof thespruewas taken intoaccount the timeto fillincreased.Theoverallfillingtimeinthesimulationwasstilltoolow,butincreasing.Addingahotrunnertorepresentthemoltenmaterialpresentbeforethenozzleaddedmuchmoretime,asthislargecushionofmaterialisbeingcompressed,withtheeffectofslowingdownthefilling.Theseaspectsofthesimulationsarediscussedindetailinthenextparagraph.

3.3 SetupofinitialandimprovedsimulationsThe first simulation conducted showed large time differences compared to the experiments, as thesimulationfilledtoofast.Thiswillbereferredtoasthereferencesimulation,andthiswasimplementedwiththementionedramspeed,thetemperaturesandthedifferent(packing,cooling)timesspecifiedintheexperimentalsetup.Thesizeandthegeometryofthecushioninfrontofthesprueanduptothenozzlecanhaveasignificantimpactonthefillingsimulationofthecavityandmustbetakenintoaccount[12][13][14].Thisisofparticularimportancewhenthevolumeofthecushionislargerthantherunner/cavitysystemtobemolded.This is indeed the case for these experiments conducted in thepresent study, as the screwdiameterisof35mmaspreviouslymentioned.Thestrokelengthofthepistonwasaround8.5mmresultinginaninjectionvolumeofapproximately8.2cm3,whichissignificantlylargerthanthevolumeofcavityandrunnerwhichisaround3.2cm3.Thismeansthatallofthemoltenmaterialstandinginfrontofthescrewwillbesubjectedtoacertaindegreeofcompressionduringtheinjectionphase,meaningthattheflowratewillslowlyincreaseinthenozzleduetothemeltcompressibility.Thiseffectresultsinahigherfillingtimeinexperimentsthaninthereferencesimulation.

Tocircumventthisproblem,thefollowingdifferentsimulationsetupswereemployedtoaccountforthecompressibilityofthemeltvolumeinthecushionandinthenozzle:

1. Referencesimulation,usingtheramspeedprofilefromthemachineandthegeometryofthecavityandrunner;

2. Addedaccelerationoftheramspeed;3. Addednozzleandmachinegeometryrepresentedasahotrunnertoaccountfortheeffectofthe

compressiblemoltenmasslocatedbeforetherunner,seeFig.4(bottom);4. Bothaddedaccelerationandnozzleandmachinegeometry;5. Sameaspoint4.,butwithslightlylowermelttemperatureinthecaseofHVPC.

8

ThesimulationswererunonanIntelCorei72.5GHz,16GBrandomaccessmemory,withatotalnumberof1millionelements.Approximately8hourswereneededtoruneachsimulation.

Simulationresultsshowedthatincludingthescrewaccelerationgavebetteroverallfillingtimecomparedtotheactualinjectiontimeintheexperiments,andbetterpredictionofthepressurecurves(seeFig.8andFig.9).

Theuseofahotrunnertosimulatethecushionofmeltisjustified,sinceahotrunnerassumesthatitisfilledwithmoltenpolymer,whichisactuallyatthetemperatureofthemelt,as ithappensinthebarreloftheinjectionmoldingmachine.Themeltmighthavedecreasedintemperaturebeforehittingthenozzle,buttheassumptionisgoodenoughforABS,duetolowsensitivitytotemperatureofitsviscosity.FortheHVPCthemelttemperaturemightbeslightly lowerthanexpected,andsincetheviscosityofthispolymerexhibitshighsensitivitytemperature(10‐15°Clowermelttemperatureisenoughtocausethistobe1.5timesmoreviscous),aslightlylowermelttemperaturewasincludedinthesimulationforHVPC.

Other effects investigatedwere transient thermal settingof themold temperature,pressuredependentviscosityandwallslip.Theywillbeexplainedindetailinthefollowing.

3.3.1 TransientthermalsimulationUsing a transient cooling simulation prior to the filling simulation, and then using the calculatedtemperaturedistributioninthemoldduringfilling, ispossible.InthecaseofABSthecoolingwaterwas17°Candlocated12mmfromthecavitysurface.Thisdoesleadtoslightlyhighertemperatureatthecavity,ofaround22°C,calculatedfromthetransientcoolingphase,andseeninthetemperaturemeasurementofthermocouple T3. The improvement was small for ABS, and it was satisfactory to run the simulationassumingsuchconstanttemperatureforthemold.ForHVPCthecoolingwaterwasat95°Candthetransientsimulationproveditstemperaturetobearound90°C.ThistemperaturewasconfirmedbythemeasurementfromthethermocoupleT3duringtheexperimentswithHPVC.

3.3.2 PressuredependentviscosityPressuredependentviscosityhasprovedtohaveaneffectonthefillingpressurerequired,andthepackingpressurelevels.ThisisshownbyPantanietal.[12]andGuerrieretal.[14]togiveahigherpressuredropthroughthemelt,givingmorerealisticpressurepredictions.TheD3coefficientfromEq.8wascalculatedby themethod presented by Pantani et al. [12] and simulationswere carried out. This did in essenceincreasethepressurerequiredtofillthecavityduetoramspeedprescribedboundarycondition.

3.3.3 WallslipTheHVPCmaterialwasobservedtonotproperlysticktothemoldwallduringinjection(i.e.thezerovelocityatwallconditioncouldprobablynotbeappliedthroughoutthewholesimulation).Hencethepossibilityofwall slip occurrencewas considered. This phenomenon could be seen in one of the high speed videosrecordedduringtheexperimentswithHVPC,inwhichacertainamountofjettingjustafterthemeltentersthegatewasvisible(seeFig.11).Awallsliptypeboundaryconditioncanbeenabledinthesoftware.Thisoptionshowedthatbyusingaverylowfrictioncoefficientof0.1,themeltwouldfillthemoldmoreeasily,aslowerpressuresarerequired,andthepolymerslidesinthe0.5mmthicksection.However,thisbehavior

9

wasnotobservedinthehighspeedvideos.Asmallimprovementinthesimulationwasobtainedbyusingahigherfrictioncoefficientof0.7,butnottotheextentasofusingalowermelttemperature.

3.4 ResultsanddiscussionTheflowischaracterizedbypassingthroughdifferentsectionsofcharacteristicgeometries.Afterthefirstpoint(i.e.ZeropointinFig.5),thefollowingpointsarethendeterminedtocharacterizethedifferentfillingsteps:

PointAispositionedincorrespondencetothefirstcornerintherunner,andthetimeisdefinedasthemomentatwhichthemeltcompletelyfillsupthecorner.

PointBisdefinedwhenthemeltisatthegate,i.e.whenitisjustabouttoenterthecavity.

PointCisindicatedaswhenthemelthasenteredthecavityandthetwofirstcornersarefilled.

PointDindicateswhenthetwomeltfrontsstarttoformaweldline,meetingafterthe0.5mmthicksection.

PointEisdefinedaswhenthemelthasflowedoverthelocationofthelastthermalsensorsT3.

Fig.5–Pointsthatindicatethetimingmeasurements.

ThesetimingsareplottedinFig.6andFig.7astheintervalbetweenthepointsandthetotaltimefromthezeropoint.Thetotaltimereflectsthetotaltimethemeltisvisibleinthehighspeedvideofieldofview,i.e.itisshorterthanthetotalexperimentalinjectiontimeof1s.AsseeninFig.6fortheABS,thereferencesimulationtendstofill fasterthanintheexperiment.Addingtheaccelerationofthescrewdoesseemtoimprovetheresultsclosetothetimingsofthehighspeedvideo,anditdoesincreasethetotalinjectiontimeclosetotherealtime.Anactual improvement isseenwhenthehotrunnergeometryforthenozzleandbarrelisincluded.Nowthetimingsarewithin10‐15msfortheintervals,and20‐25msforthetotalvideo

10

time.Theresultisthatbyusingbothaccelerationandnozzle/barrelhotrunneritispossibletoachieveagoodagreementforthehighspeedvideotimingsandthetotalinjectiontimewiththeexperiments.

Fig.6–ExperimentalandsimulatedtimingscomparedforABS

InFig.7thesameresultsarerepresentedforHVPC.ThesametrendisseenaswithABS,butthereisstilladifferenceinbothtimingswhenonlyusingtheaccelerationandthenozzle/barrelhotrunner.Thisisimprovedwhenusingalowermelttemperature,asthepolymerseemstosolidifyandblocktheflowbeforethepackingphasestarts.

Fig.7–ExperimentalandsimulatedtimingscomparedforHVPC

ThepressurecurvesfromthesimulationatthesensorlocationcanbeseeninFig.8forABSandinFig.9forHVPC.As shown, thepressurecurvesaremuchsteeper in the reference runs.When theacceleration isaddeditgivesaflatpressureincreaseinthebeginning,alsogivingalongerfillingtime.Addingthecushioncompression,thepressureincreaseissmoother,asseenintheexperiments.Finally,averygoodagreementisobtainedforABSusingbothaccelerationandthecushion,asthepressurecurveinFig.8correspondstotheexperimentalcurveshowninFig.3.FortheHVPC,usingaslightlylowermelttemperatureseemedto

11

improvetheresultsfurther,butthereisstillavisibledeviationfromtheexperimentalresults.Also,asseeninthefigureofthehighspeedvideoinTable2,thereisasolidifiedtipofthematerialcomingintotherunnerfirst, andgetting stuck there.Thiscoldplugmightmake itmoredifficult for thematerial to flow in theexperimentsthaninthesimulations.Furthermore,itseemsthattheHVPCismorepronetowallslipping,whichwasveryapparentintheexperimentsperformedathigherinjectionspeed,seeFig.11.Ascanbeseenintheimages,theHVPCshowedthetendencytoproducejettingafterthegate.

Foraqualitativecomparisonofthesimulationsandexperiments,thehighspeedvideosandsimulationsshowninTable1andTable2forthepointsA‐Eandthe‘ENDOFFILL’point(definedaswhenthemeltfillsthetwoendcorners,astheVisnotcompletelyfilledforHVPC)areconsidered.ThetimingsarenotedatthesamepointsatthoseinFig.6andFig.7.Thevisualcomparisonseemstobeingoodagreement,especiallyforABS,astheV‐shapeoftheweldlineisalmostidentical.InthecaseoftheHVPC,theV‐shapeoftheweldlineseemstobesmaller,indicatingthat ifhasslightlymoreflowabilitywithrespecttotheexperiments.Theseobservationsarealsoconfirmedwhencomparingtheexperimentaltimingswiththoseobtainedinthesimulations(seeTable3).

Fig.8–SimulatedpressurecurveforABSatthetwopressuresensorlocation,andtheboundarypressureontherunner/hotrunnerdenotedP0.

12

Fig.9–SimulatedpressurecurveforHVPCatthetwopressuresensorlocation,andtheboundarypressureontherunner/hotrunnerdenotedP0.

Fig.10–ExperimentalandsimulatedpressuresP0,P1,P2forABS(left)andHVPC(right)obtainedwiththebestsimulationsetting.‘S’denotessimulation,and‘E’experiments.

13

Table1–Experimentsandsimulationresultsofflowfrontposition,patternshapeandfillingtimewithABS.Thetimezeropointissetwhentheflowentersthefieldofviewofthehighspeedcamera.Thesimulationincludesbothaccelerationandhotrunnernozzlegeometryandmeltcushion.

EXPERIMENT SIMULATION EXPERIMENT SIMULATIONA

20ms

33ms

D

330ms

338ms

B

76ms

104ms

E

390ms

410ms

C

162ms

171ms

ENDOFFILL

536ms

478ms

14

Table2‐Experimentsandsimulationresultsofflowfrontposition,patternshapeandfillingtimewithHVPC.Thetimezeropointissetwhentheflowentersthefieldofviewofthehighspeedcamera.The

simulationincludesbothaccelerationandhotrunnernozzlegeometryandmeltcushion.

EXPERIMENT SIMULATION EXPERIMENT SIMULATIONA

36ms

37ms

D

794ms

563ms

B

122ms

107ms

E

1208ms

957ms

C

268ms

177ms

ENDOFFILL

1672ms

1176ms

15

Table3–Absoluteandrelativedeviationsofsimulatedtimeswithrespecttoexperiments.Absolutedeviationt=t(simulation)‐t(experiments);relativedeviation%=t/t(experiments)100.

SECTIONABS HVPC

ABSOLUTEDEVIATION RELATIVEDEVIATION ABSOLUTEDEVIATION RELATIVEDEVIATIONA 13ms 65,0% 1ms 2,8%B 28ms 36,8% ‐15ms ‐12,3%C 9ms 5,6% ‐91ms ‐34,0%D 8ms 2,4% ‐231ms ‐29,1%E 20ms 5,1% ‐251ms ‐20,8%

ENDOFFILL ‐58ms ‐10,8% ‐496ms ‐29,7%

Fig.11–Jettingoccurringjustafterthegateat42ms,46ms,54ms,afterthetimezeropoint.

4 ConclusionTheuseofaglassmoldisaneffectivemethodofperformingadirectvisualcomparisonofinjectionmoldedpartswithsimulations.Thiswasachievedinthepresentstudyintermsofvisualfillingpatternandalsointermsofactualtimingofthemeltfrontduringfillingandpacking.Themolddesignwithaperpendicularinjection plane to the mold opening direction presented in this work provided a convenient camerapositioning.Thermocouplesandpressuresensorswereinstalledandhelpedgettingtheoverallfillingtimescorrect,thatarenotcapturedbythehighspeedcamera,andobtainingvaluableinformationregardingthevelocity‐topressure‐controlswitch‐over,sincethisisnotvisibleinthevideorecordings.Simulationshavebeen conducted using themachine inputs and compared with timings from the high speed video andpressuresensors.Thereference(i.e.default)simulationshowedlargedeviationsofthetimingscomparedwiththeexperimentalhighspeedvideos.Thereasonwasmainlyduetothecushionsizeforthemoldingswhichislargerthanthevolumeofthepartandrunnervolume.Theothereffectwastheinclusionofthebarrelaccelerationinthemachineinputdata,whichincreasedtheinjectiontime.Withoutusingbotheffectsthesimulationwasfillingtherunnerandcavitytooeasily,resultinginalowerfillingtimethanthatseenintheexperiments.Themost accurate simulation results showedsimulationsdeviationswithin10‐30ms(relativedeviationinorderof5‐10%)fortheABSandslightlymoreforthehighviscosityPC,intherange

16

of100‐500ms(relativedeviationinorderof20‐30%)ontimingsbetweendifferentsectionsduringfilling.ThereasonfortheHVPCnotbeingasclosetotheexperimentsasABS,issuspectedbethattheviscosityismoresensitivetotemperature,itshowedsomewallslipandthatacoldplugisobservedintherunnerinthehighspeedvideos.ThepressurelevelsfromthesensorsagreeverywellforABS,andforHVPCthereisaslightdeviation,mostlyinthetimings,sincethecurvatureandthelevelsofthepressureprofilesareinfairlygoodagreementwiththeexperiments.

AcknowledgmentThisworkwascarriedoutaspartoftheInnomoldProject(InnovativePlasticProductsandMoreEnergyEfficientInjectionMoldingProcesses)supportedbytheDanishCouncilforTechnologyandInnovation.

References[1] S.Yang,S.NianandI.Sun,"FlowVisualizationofFillingProcessduringMicro‐InjectionMolding,"

Intern.PolymerProcessing,vol.4,pp.354‐360,2002.

[2] S.C.NianandS.Y.Yang,"FlowVisualizationofFillingwithAidofColoredBilletsDuringImpactMicro‐InjectionMolding,"Intern.PolymerProcessing,vol.4,pp.402‐407,2004.

[3] X.HanandH.Yokoi,"VisualizationAnalysisoftheFillingBehaviorofMeltintoMicroscaleV‐GroovesDuringtheFillingStageofInjectionMolding,"Polym.Eng.Sci.,vol.46,pp.1590‐1597,2006.

[4] S.HasegawaandH.Yokoi,"VisualizationofMelt‐FlowBehaviorInsidetheRunnerinUltraHighSpeedInjectionMolding,"Intern.PolymerProcessing,vol.5,pp.464‐472,2006.

[5] G.S.LayserandJ.P.Coulter,"EffectofProcessingParametersonFlowHesitationwithGlass‐FilledReinforcedThermoplastics:ExperimentalInvestigationwithDirectVisualization,"ANTEC,pp.342‐348,2008.

[6] R.Spares,B.R.WhitesideandP.D.Coates,"HighSpeedThermalImagingofMicromoulding,"ANTEC,pp.2570‐2573,2009.

[7] B.Whiteside,R.SparesandP.Coates,"Rheo‐OpticalMeasurementsinMicromoulding,"ANTEC,pp.1803‐1810,2009.

[8] H.YokoiandD.Yoshida,"Ultra‐High‐MagnificationVisualizationofMeltFillingBehaviorsinLine&SpaceReplicationMoldingUsingaMicroscopeBuilt‐inMold,"PPS,vol.27,2011.

[9] A.S.Yanev,G.R.DiasandA.M.Cunha,"DirectVisualizationofConventionalInjectionMolding,"ANTEC,pp.1097‐1101,2007.

17

[10]A.S.Yanev,G.R.DiasandA.M.Cunha,"VisualizationofInjectionMouldingProcess,"MaterialsScienceForum,Vols.587‐588,pp.716‐720,2008.

[11]P.Dvorak,T.BarriereandJ.C.Gelin,"Directobservationofmouldcavityfillinginceramicinjectionmoulding,"JournaloftheEuropeanCeramicSociety,vol.28,pp.1923‐1929,2008.

[12]U.Vietri,A.Sorrentino,V.SperanzaandR.Pantani,"ImprovingthePredictionsofInjectionMoldingSimulationSoftware,"Polym.Eng.Sci.,vol.51,pp.2542‐2551,2011.

[13]G.Tosello,F.CostaandH.Hansen,"MicroInjectionMouldingHighAccuracyThree‐DimensionalSimulationsandProcessControl,"ProceedingsofPolymerProcessEngineering11–EnhancedPolymerProcessing,pp.413‐435,6‐7December2011.

[14]P.Guerrier,G.ToselloandJ.H.Hattel,"AnalysisofCavityPressureandWarpageofPolyoxymethyleneThinWalledInjectionMoldedParts:ExperimentsandSimulations,"Proceedingsof30thInternationalConferenceofthePolymerProcessingSociety(PPS‐30),Cleveland,OH,p.5pages,8‐12June2014.