ELSEVIERJournalofMaterialsProcessingTechnology55(1995)408-416MaterialsProcessingTechnologyAnalysisofdiedesignsforthestampingofanautomobilerearfloorpanelFuh-KuoChen*,Jia-HongLiuDepartmentofMechanicalEngineering,NationalTaiwanUniversi,Taipei,Taiwan,ROCReceived10October1994IndustrialsummaryThestampingprocessofmanufacturingaone-piecerearfloorpanelofapassengercarhasbeeninvestigatedinthepresentstudy.Ananalysisoftheoriginaldiedesign,inwhichasplitdefectoccurredatthedrawn-cupwall,wasperformedusingcircle-gridanalysisaswellasthe3-Dfinite-elementmethod.Thesplitdefectisduetothelargeareaofsheetmetalundertheblank-holderthatlimitstheflowofthemetaltowardsthecuparea.Anoptimumdiedesign,whichconsistsofaseparatediefaceandawedgemechanismmountedinthelowerdieframe,wasproposedtoprovideadditionalmetalforthecupareaandtoeliminatethesplitdefectwithoutaddinganextraoperation.Thisoptimumdiedesignwasvalidatedbytheresultsofcircle-gridanalysisperformedforthefirstandseconddrawoperationsandalsobythegood-qualitypanelsproduced.Keywords:Stampingdies；Rearfloorpanel；Splits；Circle-gridanalysis.1.IntroductionSplitsareamongstthemajordefectsoccurringcom-monlyinthestampingprocess.Alotofresearchefforthasbeenmadetoinvestigatethecausesofandsolutionstothesplitproblemduringthelastfewdecades1-4,themethodsusedincludingforming-limitanalysisandthefinite-elementmethod.SinceKeelerandBackofen5-firstintroducedtheconceptofforming-limitdiagrams(FLDs)in1963,theyhavebeenusedwidelyintheana-lysisofsheetmetalforminginpressshops.TheFLDsindicatethestrainswhichleadtofailureandthusprovideausefultooltodetermineiftheformingprocessislikelytobepronetosplitting,whilstthefinite-elementmethodcancalculatethestraindistributionsinthestampedpartsaccuratelyandthuspredictifthesplitdefectislikelytooccur.Ingeneral,thesolutiontoproblemsofsplittingistoprovidemoremetaltothecriticalareabeforethemajordrawingprocessstarts.Thiscanbeachievedeitherbydecreasingtheblank-holderpressureorimprovingthelubricationconditions,butthemoststraight-forwardmethodistoaddanextraoperationsolelyforthepur-poseoffeedingmoremetalintothecriticalarea.*Correspondingauthor.0924-0136/96/$15.001996ElsevierScienceS.A.AllrightsreservedSSDI0924-0136(95)02038-NHowever,theextraoperationincreasestheproductioncostbyaddingonemoresetofdiesandadditionalman-power；andhenceinreality,itshouldbeavoided.Inthepresentstudy,anoptimumdiedesigninwhichaseparatediefaceandawedgemechanismmountedinthelowerdieframeisproposedtoeliminatetheoccur-renceofasplit-defectinthestampingprocessforaone-piecefloorpanelofapassengercar.Thespecialdiefaceandwedgemechanismweredesignedtoprovideaddi-tionalmetalforthecriticalareawherethesplitdefectoccurred,withoutaddinganextraoperation.Bothcircle-gridanalysisand3-Dfinite-elementsimulationswereperformedtoanalyzethissplitdefect.2.ProblemdescriptionThecommondesignforarearfloorpanelofapassen-gercarisusuallyatwo-piecetype,namely,weldingtwostampedpiecestogether,asshowninFig.1.Thetwo-piecetypedesignischosenmainlyduetothedifficultyencounteredinstampingaone-piecerearfloorpanel,inwhichasplittendstotakeplaceatthewallofthedeeplydrawncupusedforstoringthesparetire,asshowninFig.2.Theoccurrenceofthesplitisattributedtothesubstantialdistancebetweenthecupwallandonesideoftheblank-holder,asshownbylineA-BinFig.3,thatF.-K.Chen,J.-H.Liu/JournalofMaterialsProcessingTechnology55(1995)408416409)Fig.1.Atwo-piecerearfloorpanel.splitFig.2.Asplitoccurringatthecupwall.Fig.3.Thelocationofthedrawncupinthesheet-blank.restrictsthemetalundertheblank-holderfromflowingintothecuparea.Whilstinthetwo-piecedesignthisdistanceismuchshorterandsufficientmetalcanfloweasilyintothecuptopreventtheedgeofthecupfromsplitting,duetocost-effectiveconsiderations,aone-piecerearfloorpanelisalwaysdesired,sothatthesplitprob-lemmustthereforebeovercome.Theoriginalprocedureinthepressshopforthepro-ductionoftheone-piecerearfloorpanelconsistsoffouroperations:firstdraw,seconddraw,trimming,andflang-ing.Thefirst-drawoperationwasdesignedonlytopro-duceacupshape,asshowninFig.3.Astortheribsaroundthecup,thesewereformedinthesecond-drawoperation.Likemoststampingprocesses,themaindefor-mationoftherearfloorpaneliscompletedinthefirst-drawoperation.Theconventionaldrawprocessallowsthepunchtopullmoremetalintothediecavityfromtheblank-holder.Tofacilitatemetal-flow,nodrawbeadswereemployedontheblank-holdersurface.However,duetothelargedrawdepthandthegeometricdifficultymentionedabove,asplitwasstillfoundatthecupwallnearthebottomafterthefirst-drawoperation,asshowninFig.2.Thelocationofthesplitdefectindicatesthattheconsiderabledistancebetweenonesideofthecupwallandtheblank-holderdoespreventthemetalfromflow-ingtowardsthecup.Someeffortshavebeenmadetohelpthemetaltoflowtowardthecuparea.The,attemptofdecreasingtheblank-holderpressureledtomorewrinklesattherootofthecupareabutwithouteliminat-ingthesplit.Improvementofthesheetmetalqualitywasalsoproventobeinvain.Extremecaretakenwiththelubricationconditionscaneasethesplitproblem:how-ever,thisisnotcost-effectivefortheprocessofmassproduction.Alsothelargequantityoflubricationoilusedinthestampingprocessmaypollutetheshopfloor.Hence,themostefficientwaylefttosolvethisproblemistoprovidemoremetalforthecupareabeforethepunchstartstoformthecup.Toachievethisend,modificationoftheblank-holdersurfacetoformabetterbinder-wrapshapewhichprovidesmoremetalaroundthecupareawasconsidered.Thebinder-wrapisthedeformedshapeofthesheet-blankattheclosureoftheblank-holders.However,duetothesamegeometricalreason,i.e.theconsiderabledistancebetweenthecupandonesideoftheblank-holder,theoptimumblank-holdersurfaceisnoteasytoobtain.Finally,aseparatediefacedesignedforthefirst-drawoperationaidedbyaspecialwedgemechanismmountedinthelowerdieframeprovidedmoremetalforthecupareaandenabledtheproductionofsoundprod-uctswithoutsplitdefects.3.AnalysisoftheoriginaldesignTheproblemofsplittingisusuallyrelatedtothestraindistributioninthecriticalarea.Thestraindistributioninanycrosssectionoftheformedpartisdeterminedbytwofactors:oneistheamountofmetal-flowresultingfromdraw-inovertheblank-holder；theotheristheamountofstretchcreatedbythecontactsbetweenthepunchandthedie6.Toquantifytheeffectofthetoolinggeometryonthemetalflow,theoriginaldesignwasanalyzedbycircle-gridanalysis(CGA)andthefinite-elementmethod(FEM).410F.-K.Chen,J.-H.Liu/JournalofMaterialsProcessingTechnology55(1995)408-40.20.1TRUESTRESS(GPa)0TRUESTRAIN0.40.5Fig.4.Thestress-straincurveforthesheet-blank.3.1.Circle-gridanalysisCirclegridanalysishasbeenusedwidelyinthepressshoptomeasurethestraindistributionsandthustoenableananalysisoftheformabilityofthesheetmetalbyplottingthemeasuredstrainsontheforming-limitdia-gram.Thecirclegridshaveamajoradvantageovertheothertypesofgrid,suchassquaregrids,sincetheydonothaveanypreferredorientations.Thisadvantageliesinthatthedeformationofthecircleswillresultinellipses,thetwoprincipaldirectionsbeingdisplayedclearlybythemajorandminoraxes.Themagnitudesoftheprinci-palstrainscanthenbecalculatedfromthemeasurementsofthelengthsofthemajorandminoraxes.Byplottingthesemeasuredmajorandminorstrainsontheforminglimitdiagram,theseverityofareasofaformedpartcanbeevaluated.Inthepresentstudy,aproductionfloor-panelmadeusingtheoriginaldiedesignwasfirstanalyzedwithcircle-gridanalysis.Thesteelsheetusedforproductionis0.7mmthickandofDDQquality,withmaterialproper-tiesasshowninFig.4.Thecorrespondingforming-limitdiagramforthismaterialprovidedbythesteelsupplierisshownbythesolidcurveinFig.5.Anyareaofaformedpartwhichisdeformedwiththestrainslocatedonorclosetothiscurvewilltendtosplit.Inpractice,theforming-limitcurveshifteddownby10%,asshownbythedashedlineinFig.5,isusedasthedesigncurve.Theareaabovetheforminglimitcurveiscalledthefailurezone；theareabetweentheforming-limitcurveandthedesigncurveistermedthemarginalzone；andtheareabelowthedesigncurveisnamedthesafezone.Ingeneral,thestraindistributionsinanyareaofaformedpartshouldfallinthesafezonetomakethestampingprocessstable,astampingprocessbeingsaidtobestableifitislesssensitivetoprocessvariations.Beforebeingstamped,thecriticalareaofthesheet-blankwasimprintedwith5-mmdiametercircleswith6-mmspacingbetweentheircentres.Toimprintthecirclesthecriticalareaofthesheet-blankwasfirstcleanedusingaspecialcleaner,thenastencilwiththecorrectgridpatternwasplacedinpositiononthepart.Usingtheelectrolyteasconductor,theareacoveredbythestencilwasmarkedwiththegridpatternbyanetchingprocess.Topreventthemarkedareafromrust-ing,apieceofclothsaturatedwithcleanerwasusedtowipeoffexcesselectrolyteandresidualoxidesinthemarks.Afterbeingstamped,asplitwasfoundonthecupwallneartothetop,asshowninFig.2.Themajorandminorstrainsofthedeformedcirclesaroundthesplit,asshowninFig.6,weremeasuredandplottedontheforminglimitdiagram,asshowninFig.7.Ascanbeseeninthisfigure,themeasuredstrainsareeitheraboveorclosetotheF.-K.Chen,J.-H.Liu/JournalofMaterialsProcessingTechnology.55(1995)408-41641190858075706555#5045_4003503r3025Failurezonefezone2015105-30-25-20-15-10-50510152025303540MinorStrain(%)Fig.5.Theforming-limitdiagramfor0.7mmDDQsteelsheet.9085807570656o55r-_5O-,-4540.3530252015105J-30-25-20-15-10-50JII!II510152025303540MinorStrain(%)Fig.7.Measuredstrainsaroundthesplit.Fig.6.Deformedcirclesonthesheet-blank.forming-limitcurve,andthefailureisobviouslyduetostretching,sinceboththemajorandminorstrainsarepositive.Itisalsonotedthatthestrainsareveryclosetotheplane-strainfailuremode,i.e.,closetotheaxisonwhichtheminorstrainiszero.Theresultsofthecircle-gridanalysisindicatethattheoriginaldesignisveryunstable.TheFLDalsoindicatesthatthemajorstrainistoolarge:thisisconsistentwiththepresentauthorsopinionthattheconsiderabledis-tancebetweenthecupandonesideoftheblank-holderlimitstheflowofmetaltowardsthecuparea,resultinginthelargestrains.Asdiscussedintheprevioussections,themosteffectivemethodofdecreasingthemajorstrainistoprovidemoremetalforthecuparea.3.2.FiniteelementanalysisTohelpfurtherunderstandi