DOC

Stresses

By Sam Martin,2014-02-18 23:33
8 views 0
StressesStres

    Stresses

    Stressesdistributionafterplaningandcuttinginweldedplates55

    Stressesdistributionafterplaningandcuttinginweldedplates

    ZhouJianlai,LiuShuchunandYaoChuanwei

    周建来,刘树春,姚传维

    AbstractToforecastthemodificationandevolutionofaresidualstressfieldinabutt

    weldedplate,theresidualstressfield

    wasnumericallyevaluatedafiniteelementmethodcode.Theplatewasoriginatedbyweldingandsufferederc

    formingmachining.Theresidualstressfieldwastakenasapre

    stressconditionforthenextcuttingsimulations.Residual

    stressdistributionermachiningwasthencomparedtothatintheinitialresidualstressfield.Numericalresultshavebeen

    comparedtoexperimentalmeasurements.Thatshowsthepotentialityaswellasthelimitationsofnumericaltechniques.Three

    m4orcontributionsaresummar~edasfollows:Longitudinalresidualstressdistributioninweldedplatesisdeeplychangedby

    mechanicaltooling;Planingandcuttingincreasespeakvaluesoftransversalresidualstressandingeneralintroducessome

    unevennessindistributionalongX-direction;Comparisonofexperimentalandnumericalvaluesisgenerallysatisfactoryfor

    longitudinalresidualstress.

    Keywordsresidualstress,welding,planing,finiteelementmodeling

    0Introduction

    Duringweldingthematerialundergoesacomplex

    thermalandmechanicalhistory,andtheresultingassem

    blyinheritsamemoryofthisprocessintheformofmicro

    structuralandcompositionalchangesandresidualstress distribution.Andresidualstressesdevelopduringmost manufacturingprocessesincludingmaterialdeformation, heattreatment,machiningorprocessingoperationsthat transformtheshapeorchangethepropertiesofamateria1. Theincreaseofcomputationalcapabilitiesallowsa morepreciseandpunctualsimulationofthethenno-me?-

    chanicalprocesswhichoriginatesresidualstressinme- chanicalcomponents.Simulationsoftheweldingprocess andevaluationofresidualstressarethemainobjectiveofa largenumberofpapers.Inparticular,theindicationsre

    portedinreference[1]areveryclearandadoptedforthe numericalmodelpresentedinthispaper.

    Bynowreliablepredictionofmaterialstructureand residualstressstateattheoutcomeofaweldingprocessre

    mainselusive,itsusebeingprimarilylimitedtoqualitative identificationof,forexample,themostfavourablecondi

    tionsforproducingaweldwithlowerresidualstressand distortion.Inaddition,machining'inducedresidualstres? seshavebeenstudiedsinceHenriksen'sfundamental work.Recently,someexperimentalworkandnumerical modelshavebeenpresentedforevaluationandprediction ofresidualstressinducedbymachiningprocess[. However,asolid,accurate,understandabkmodel fornumericalpredictionoftheevolutionandmodification thattheresidualstressfieldofabuttweldedplatesuffers

    afteracuttingprocesshasnotyetbeendevelopedtothe pointofgeneralacceptance.Furtherinvestigationisre- quiredtounderstandtheresidualstressdistributionafter machiningoftheweldedplate.

1Geometryandmodeldescription

    Theworkpiecematerialislowalloysteelmainlycon

    taining0.4wt%C,1.8wt%Ni,0.8wt%Crand0.7

    wt%Mo.Itismodeledwithaviscoplasticconstitutive modeloftheoverstresspowerlawtype:

    8=D(1)

    n

    (1)

    ZhouJianlai,LiuShuchunandYaoChuanwei,SchoolofMechanicalEngineering,HuaihaiIn

    stituteofTechnology,Lianyungang,

    222005.ZhouJianlai,Correspondingauthor,Email:a177654444@126.con. 56CHINAWELDINGVo1.

    19No.3September2010

    for?0,where

    p

    istheeffectiveplasticstrainrate,

    isthecurrentflowstress

    ,0istheinitialyieldstress.and

    Dandmarematerialparameters.Thisrate.dependent powerlawishighlysuitableforhighstrain.rateaDplica. tionslikehighspeedmetalcutting.

    Standardconstantva1.

    uesareusedforotherphysicalproperties.

    Theenon0us

    heatgeneratedduringthemetalcuttingprocesswilllocallv alterthematerialpropertiesoftheworkpiecemateria1.S0. temperature?dependentmaterialproperties(e.g.elastic constants,initialyieldstress,andthermalexpansionc0ef- ficient)areusedE.

    NumericalanalysisisperformedbyABAQUSand

    QUICKTHERMontheweldedjointsobtainedfromplates ofthreedifferentthicknessesshowninFig.1. 12345

    ff00

    n

    

    

    WeldfneII

    ,.-_?-__?-y

    Fig?1Geometryofweldedplatesandtransversaljoints Themodelexploitstheproblemsymmetryastothe weldaxis,whichpermitstoconsiderahalfplateon1v. Themodelusessolid8.nodelinearelements.Afinerdis

    cretizationclosetotheweldseamhasbeenperformed. Thewidthofthesmallerelementisch0sen0nthebasis0f themeshsensitivityanalysiscarriedoutinpreviouswork. Themaximumvalueoflongitudinalresidualstressindif- ferentmodelsisreportedagainstthereciprocalvalueofthe elementlengthintheplatethicknessdirection. Residualstressfieldhasbeenevaluatedfirstlv0fa11 asaresultofaweldingprocessfor8.12and16mmplate thicknesses.Secondly,eachweldedplateissubjectedtoa planingprocessthatremovesamateriallayer2mmthick. Finally,eachplateiscutinordertoobtain10transversa1 weldedjointswith80mmwidth.Theselattermachinings produceachangebothinmagnitudeandindistribution0f residualstressfield.Numericalsimulationspermittofo1. 10wtheresidualstressstateevolutionfrom?weldedplates uptofinalgeometry.

    2Simulationsandexperiments

Thedimensionofamodelis300mmx800mm×trt

    :8,12,16mm).Temperaturedependencies0fmaterial properties,suchastheyieldstressor,Young'smodulu E,thermalexpansioncoefficientandthea1conductivi

    tyKwereconsidered.

    Theweldingprocesssimulationistreatedasanun. coupledthermalandmechanicalanalysis:firsttemperature fieldiscalculated,

    thenstressanddisplacementsarede.

    ducedfromtheseresults.LinearelementshavebeenDre

    ferredtohigherorderelements.Realisticboundarycondi. tionsoftheplateisconsideredinthenunerica1mode1dur_ ingwelding.Inparticular,azerovalueisimposedtothe displacementintheZdirection(thicknessdirection)ofa11 theedgeBnodes.Thepossibilityofarigidmotionofthe plateisavoidedbyintroducingafixedendonedgeA (Fig.1).

    Numericalsimulationconsistsofthreephases: (1)Evaluationoftemperaturefieldintheplatesbv meansofathermaltransientanalysis.Gasmetalarcweld. ingissimulatedbyauniformheatfluxcomprehensive0f arcthermalenergyandmetaldropsthermalenergytrans. ferredfromthefillerwire.Themotionoftheheatsource alongtheweldcordisobtainedbymeansofaFORTRAN subroutine.Weldingparameterschosenforthisanalysis arecharacteristicofgastungstenarcwelding;(2)Appli.

    cationofthethermaloutputinordertoobtainresidual stressfield;(3)Removalofboundaryconditi0nsused duringtheweldingprocessexceptrigidbodyboundaries. Fig.2showsaschematicdiagramoftheorthogonal

    metalcuttingprocess,inwhichacontinuouschipisbeing takenofffromtheworkpiecebyacuttingtoolthatismov. ingrelativelytotheworkpiecewithaconstantvelocitv. Inordertomodelchipseparationandtreatfrictional interactionsinthetoolchip-workpieeesystem,threecon.

    tactpairsaredefined,asshowninFig.2.ContactPair1 definesthecuttingpath,

    wherethetwocontactsurfaces

    arerepresentedbytwosetsofnodes(oneoneachsur- face)thatar.pairedandbondedtogether.Whenthechip separationcriterionissatisfied,thecontactnodepairim. Stressesdistributionafterplaningandcuttinginweldedplates57 mediatelyaheadofthetooltipisseparated,enablingthe tooltoadvanceincrementally.Asthetoolbreaksthecon

    tactnodepairs,materialsformingthechip'sinnerface willmoveintotheregiondefinedbyContactPair2,and thoseformingthefinishedworkpiecesurfacewillmoveinto theregionofContactPair3,asillustratedinFig.2.While ContactPair2modelsthefrictionalinteractionbetweenthe chipandtool'srakeface,ContactPair3isusedonlyto maintaintooltipcontactingwiththenewlycutsurfaceof theworkpiece.

    StressstateII

    ————L.

    ?'!r\lCacfcpai.:/.

    一二.'

    /rr0.lfip\Di

    stanceaheadoftoo1tipICoactDair3

    Fig.2Aschematicdiagramofthemetalcutting

    modelwithcontactpairs

    Thefiniteelementmeshateverylayeriscomposedof 1000four-nodeplanestrainelements.Inordertomodel expectedlargedeformationinthechip,themeshforthe chiplayerisfinerthanthatfortheworkpiece.Sincethe .

    workpieceissufficientlylonginthecuttingdirectionfora-- chievingasteadystatesolution(ignoringanytransient effectsatthestartandtheendofthecuttingsimulations), theleftandfightendsoftheworkpieceboundaryarere

    strainedinthecuttingdirection.

    Inmetalcuttingsimulations,chipseparationalong thecuttingplanetakesplacewhenthestressanddeforma. tionstatesinasmallregionaheadofthetooltipsatisfya certainchipseparationcriterion.Itisworthnotingthatthe studybyHuangandBlackhasshownthatthegeometryof thechipandthedistributionofstressandstrainfieldsare notverymuchinfluencedbytheuseofaparticularchip separationcriterion_4J.Inthepresentstudy,acritical stresscriterionisusedtogovernchipseparation.Accord

    ingtothiscriterion,chipseparationoccurswhenthestress stateatacertaindistanceaheadofthetooltipreachesa criticalcombination.Mathematically,thiscriticalstress criterioncanbewrittenintermsofastressindexparame

    terfasgivenbelow:

    /=?(()(2)

    whereor=max(or2,0),andaretheshearandnor-

    realstresscomponentsofthestressstateataspecifieddis

    taneeinfrontofthetooltipalongthecuttingpath,as showninFig.2,and,and,arethefailurestressesof

thematerialunderpuretensileandshearloadingcondi

    tions,respectively.ChipseparationOccurswhenthestress indexfreachesthevalueof1.0atoneelementlength (approximately50.8Ixminthisstudy)aheadofthetool tip.Forthematerialusedinexperiments,thefailure stressesaretakentobe=948MPaandTf=f,=

    548MPa.

    Inametalcuttingprocesslocalheatingarisesbecause ofenergydissipationduetoplasticworkinthechipand workpieceandduetothefrictionalworkalongthetool- chipinterface.Inhighspeedcutting,theheatgenerated hasnotimeforconductionandtheresultingtemperature riseisusuallyconsideredtotakeplacelocally.Underthe aboveadiabaticheatingconditions,thelocaltemperature rise,AL,inducedbyplasticworkinatimeintervalAt, canbewrittenas

    ?ore~p

    t-(3)

    where.istheeffectivestress,Jtheequivalentheatcon. versionfactor,cthespecificheat,Pthemassdensity, and叩口thepercentageofplasticworkconvertedintoheat (=90%inthisstudy).

    Similarly,thelocaltemperatureriseAcausedby frictioninatimeintervalAtcanbedeterminedfrom ATS

    .

    (4)

    58CHINAWELDINGVo1.19No.3September2010

    whereistheshearstressatthecontactpoint,sisthe slipvelocity.Thecoefficientstandsforthefractionof

    frictionalworkconvertedintoheat,whichistakenas1.0 inthisstudy.

    Toobtainfivejointswith80mmwide(Fig.1),the cuttingprocessisalsosimulatedbyremovingsurrounding elements.

    3Resultsanddiscussion

    3.1Longitudinalresidualstress

    LongitudinalresidualstressdistributionalongtheX

    directionisreportedinFig.3fortheas-weldedplate,th. planedplateandthecutspecimen.Thegraphsarere. ferredtoanidenticalsection,identifiedbythecoordinate Y:360mm.whichcorrespondstothemiddlesectionof transversaljoint5.Besides,residualstressesarereported forthenodesonthephysicalsurfaceoftheplanedplate andtransversaljoint5,whilethenodesat1mmdepthare c0nsideredinthecaseoftheweldedplate.Thisway, stressdistributionsarereferredtothesamenodesforall thethreeconditions.Thenumericalresultsofthewelded plateshighlighthowthevariationofresidualstressinthe upperlayersupto2mmdepthispracticallynegligible. Asfarasresultsareconcerned,theas-weldedplates reachvieldstressontheweldcordandstressdistribution issimilartotheclassicalexpectedonel6j.

    Theplatethick-

    nessseemsnottoinfluencesignificantlybothstressdistri- buti0nandDeakvalues:onlythe16mmasweldedplate

    sh0wsalowermaximumstress,butinthiscasethesecond passoftheweldingprocesshastheeffecttopartiallyrelax residualstress.

Mechanicaltoolingaltersthepreviousstressdistribu

    tion:peakvaluesoflongitudinalresidualstressarere'

    ducedbvabout50%andtheyarelocalizedintheheataf-

    fectedzone.Stressredistributionresultingfromplaning

    has_arelevanteffectontheweldcord,whichisnowsub

    iectedtocompressivestress.Finally,cuttingoftransversal

    joints,startingfromthewholeplate,doesnotalterstress

    distributionbutitfurtherreducespeakvalues. 3.2Transversalresidualstress

    TransversalresidualstressdistributionalongtheX

    directionisreportedinFig.4fortheasweldedplate,the 500

    400

    300

    200

    100

    0

    -

    100

    -

    200

    500

    400

    300

    200

    1O0

    0

    -

    100

    -

Report this document

For any questions or suggestions please email
cust-service@docsford.com