DOC

Numerical_0

By Susan Hall,2014-07-24 00:41
7 views 0
Numerical_0Numeri

    Numerical

1270Huangeta1./JZhejiangUnivSciA20089f9J:12701276

    ;JournalofZhejiangUniversitySCIENCEA

    ;ISSN10093095(Print);ISSN18621775(Online)

    ;wwwzjueducn~zus;wwwspringedinkcom

    ;E-mail:jzus@zjueducn

    ;NumericalsimulationandptimizatiqdesigoftheEGRslmulationandoptimizationdesit~noltheER

    ;C0olerlnVeIiicle’?…

    ;Yu—qiHUANG”

    ;,

    ;XiaoliYUGuodongLU,

    ;(tPowerMachineryandVehicularEngineeringInstitute,ZhejiangUniversity,Hangzhou31002~China)

    ;(eZhejiangYinlunMachineryCo.,Ltd.,Tiantai317200,China)

    ;Email:huangyuqi@zju.educn;yuxl@zju.edu.ca

    ;ReceivedMar.26,2008;revisionacceptedMay22,2008

    ;Abstract:TheEGR(exhaustgasrecirculation)techniquecangreatlyreducetheNOxemissionofdieselengines.especiallywhen

    ;anEGRcoolerisemployed.Numericalsimulationsareappliedtostudytheflowfieldandtemperaturedistributionsinsidethe

    ;EGRcooler.ThreedifferentmodelsofEGRcoolerareinvestigated.amongwhichmodelAisatraditionalone.andmodelsBand

    ;Careimprovedbyaddingahelicalbameinthecoolingarea.InmodelsBandCtheentrydirectionsofcoolingwateraredifferent.

    ;whichmostlyinfluencestheflowresistance.TheresultsshowthattheimprovedstructuresnotonlylcagthentheflowDathofthe

    ;coolingwater,butalsoenhaneetheheatexchangeratebetweenthecoolandhotmedia.Inconclusionwesuggestthattheimproved

    ;structuresaremorepowerfulthanthetraditionalone.

    ;Keywords:Exhaustgasrecirculation(EGR)cooler,Computationalfluiddynamics(CFD),Shellandtubeheatexchanger,

    ;Helicalbame

    ;doi:10.1631/jzus.A0820223Documentcode:ACLCnumber:TK414.2+12;0357.5 ;INTR0DUCT10N

    ;Thedieselengineiswidelyusedinmodernve

    ;hicles.Unfortunatelyitisoneofthemaiorgaspoilu.

    ;tionsourcessincethetoxics,suchasnitrogenoxides

    ;(NOx)andotherpollutants.initsexhaustgascause

    ;adversehealtheffects.Itiswellknownthatdecreas

    ;ingthegastemperatureresultsinreductionsinNOxin ;theexhaustgases.Inordertorestraintheformationof ;pollutants,ahieciencyaircoolerwasdeveloped

    ;tocoolthegasofdieselengines(Grunenwaldeta1., ;2001;Pantoweta1.,2001;Stolzeta1.,20011. ;Consideringitscomplicatedandhostileworking ;conditions,anEGRcoolermustbecompactwiththe ;followingfeatures:smallvolume,steadyperformance, ;easytocleananddurableinthehightemperatureand ;corrosiveenvironment.Becauseofthat,theshelland

    ;tubeheatexchangermadeofstainlesssteelisexten

    ;lCorrespondingauthor

    ;sivelyadoptedforcoolingtheexhaustgasbefore ;mixingwiththechargeintheintakesystem. ;Muchliteratureisavailableontheoptimization ;oftheshel1.and.tubeheatexchanger.Shellsidebaf- ;tiesOfdifrerenttypeshavebeenwidelyusedto ;promoteheatexchangeforseveraldecades.Awell ;designedbafilecouldnotonlyimprovetheflowpat

    ;ternbutalsostrengthentheheatexchangeeffectively ;withlesspunishingpressureloss.LutchaandNem

    ;canskyf1990)foundthatinasingletubeheatex

    ;changer,ahelicalbafilecouldforcetheshellside ;flowtoapproachaplugflowcondition,whichin

    ;creasedtheaveragetemperaturedrivingforce.The ;flowpaRernsinducedbythebaffiesalsomadethe ;shellsideheattransferincreasesignificantly.Stehlik ;ef,.fl994)conductedacomparativeinvestigationof ;correctionfactorsbetweensegmentalandhelical ;bafiles,andprovedthesuperiorityofhelicalbafiles ;bydataanalysis.Pengeta1.(2007)presentedanex

    ;perimentalstudyoftwoshellandtubeheatex

    ;changersusinghelicalandsegmentbamesalterna

    ;

    ;Huangeta1./dZhejiangUnivSciA200899):127o_1276

    ;tively,theresultsindicatingthattheonewithcon

    ;tinuoushelica1bafes1eadstonearly10%increasein

    ;heattransfercoefflcientcomparedtoanotherone ;usingconventionalsegmentalbafflesforthesame ;shellsidepressuredrop.Nasreta1.(2007)investi

    ;gatedtheflowpattemofashel1.-and..tubeheatex.- ;changerwithhelicalbaes.anddevelopedtherela

    ;tionshipbetweenthearea,theheattransfercoe~cient ;andpressuredropbasedupontheconceptofarapid

    ;designalgorithm.Basedonthestudy,thehelicalbar- ;tiesareaproperreplacementforthesegmentalbaffles ;inshellandtubeheatexchangers.Theycouldre. ;movemanyofthedrawbacksofsegmenta1baes.

    ;Theshellandtubeheatexchangerwithhelical

    ;baffiesalsoshowedsignificantimprovementinthe ;foulingbehaviorduringoperation.Mastereta1.(2003) ;demonstratedthehelixchangeroptioninreducingthe ;velocivydependentfoulingofheatexchangers.The ;helixchangerheatexchangerwhenappliedintypical ;foulingservicesprovedtobeeffectiveinreducingthe ;foulingrates.

    ;Inconclusion.thehelicalbafehasseveralad

    ;vantages:(1)improvementofshellsideheattransfer; ;(2)lesspressuredropforagivenmassflowrate;(3) ;reductioninbypasseffectsinshellside;f41decrease ;infoulinginshellside;(5)preventionofbundlevi

    ;bration.

    ;Duetoitsoutstandingperformancethehelical ;bafehasbeenintroducedintotheEGRcoolerre

     ;cently.Pastresearchfocusedoneitherexperimenta

    ;tionwithdifferentgeometricdesignorcomparative ;analysisbetweenmeasuredresultsandevaluation.In ;thepresentresearchthreedesignmodelsbasedon

    ;EGRcoolersarepresented.Commercialcomputa

    ;tionalfluiddynamics(CFD)codesbaseduponthe ;finitevolumemethodisusedtomakethesimulation. ;Thenthenumericalresultsaretabledandcompared. ;Theflowingmechanismandtemperaturefieldsinside ;areanalyzedtoexplaintheoptimizedphenomenonin ;theheatexchanger.

    ;M0DELDESCRIPT10N

    ;ThepresentEGRcoolerconsistsofseventubes, ;twoclapboardsandashell,etc.(detailsareshownin ;Figs.1and21.Theexhaustgasflowinthetubeside ;conversetoy-axisandthewaterascoldmediumare ;1271

    ;curtainedoffbyatubewallandclapboards.A11the ;tubesandshellsinthethreemodelshavethesame ;size.Derivedfromthenoveldesign,modelsBandC ;bothhaveahelicalbaeontheshellside.Thede

    ;tailedstructureofthebaesareexactlythesame.

    ;exceptforinletandoutletsectionsofwaterflow ;whichareverticaltotheshellwallinmodelBbut

;tangentia1inmodelC.

    ;Fig.1EGRcooler(modelAandmodelB)

    ;1:Gasinletchamber;2:Shell;3:Waterinletsection;4:Gas ;outletchamber;5:Wateroutletsection;6:Clapboard;7:Tube; ;8:Helicalbaffle;9:Clapboard

    ;Fig.2ConfigurationofmodelC

    ;Aimingtostudytheflowstructureandheatex. ;change.aCFDprogramFluent6.3isintroducedto ;solvethediscretizedNavier-Stokesequationsforheat ;conductionproblemsinvolvedinasolidzone.The ;standardk-turbulencemodelwithshearflowcor. ;rectionsisusedtodealwithhighspeedturbulentflow

    ;problems.ThesecondorderupwinddifferencefUD) ;schemeisadoptedforthemomentum,energyand ;turbulenceequations.

    ;Thestandardmodelisasemiempirica1one.

    ;whichisbasedonmodeltransportequationsforthe ;turbulencekineticenergy(anditsdissipationrate ;r.TheReynoldsstressesarerelatedtothemean ;

    ;1272Huangeta1./JZhejiangUnivSciA20089f9J:1270-1276 ;velocitygradientsonthebasisoftheBoussinesq ;hypothesis.kandareobtainedfromthefollowing ;transportequations:

    ;c+c=[(+)]+pGkp6,

    ;每一+

    ;(2)

    ;Intheseequations,Gkrepresentsthegeneration ;ofturbulencekineticenergyduetothemeanvelocity ;gradients,calculatedby

    ;:+

    ;]POxj\\jOxi)(3)

    ;Thekandarecoupledtothegoverningequa

    ;tionsviatherelation

    ;=pCuk/6,(4)

    ;C11.44,C1.92,C0.09,=1.0,1.3.

    ;Theempiricalconstantfortheturbulencemodel ;isassignedinaccordancewiththerecommendationof ;LaunderandSpalding(1972).

    ;GRIDGENERATIoNANDBoUNDARYCoNDI.

    ;TIONS

    ;Thecomputationaldomainincludesboththe ;tubesideandtheshellsideregions.Aroundtheheli. ;calbameinmodelsBandCtetrahedralcellsare

    ;generated.whichareusedinparticularzonesforad. ;aptationtothecomplexgeometricalshape.Inthe ;regionsoftheinletandoutletsections,prismcellsare ;distributed.Fortheiraccuracyandstability,the ;hexahedralorprismcellsarepreferredinthegrid ;generation(Fig-3,.

    ;Inordertoestablishagrid.independentsolution, ;testsareperformed.Theresultsobtainedwithtengrid ;sizesofmodelAforacertainmassflowrateare ;showninFig.4,whichpresentapressuredropand ;temperaturedifferenceonthetubesideatdifferent ;gridsizes,correspondingtodifferentcellsnumbers.It ;iSfoundthattheresultsconvergeattheeellnumberof ;3615878,wheretheminimumsizeofmeshesiS0.3 ;1TnTI.ThisiStosaythatthesimulationbecomesgrid ;independentforthissizeofgrid;thereforeallthe ;followingsimulationsarecarriedoutwithsuchagrid. ;Ittakesabout24hforcalculationonasuper.computer ;with6CPUsinvolved.Asaresult.approximately5 ;millionmixtureelementsaregeneratedformodelsB ;andC,thesizeOfallthemeshesarrangedfrom0.3to ;0.8mm.

    ;(a1

    ;(b)

    ;Fig.3Computationalgridofmodels.(a)Clapboard ;andhelicalbaffle;(b)Tubeandhelicalbaffle ;Numberofcells(×10)

    ;Fig.4Pressuredropandtemperaturedifferencefor ;dierentnumbersofcells

    ;一一?0uuuhIp三日矗口巨=.

    ;dupI?u暑口0oJPu??uJd

    ;

    ;Huangeta1./JZhejiangUnivSciA20089f9J:1270-1276 ;Inthesimulationofmodels,boundarycondi- ;tionsarerequiredforvelocity,temperatureand ;pressure.Attheinletplaneofthewaterandgas,mass ;flow.ratesareassumedtobeconstantataparticular ;temperature.Thethicknessoftubewal1andhelica1 ;bafnearesetaszeroinmodelingduetotheirnegli- ;giblesize(inparticularwhencomparedtoother ;components).Onthetubeandhelicalbafnesurfaces. ;coupledthermalconditionsaredefinedandshell ;conductionmodelsareadoptedbyinputtingthewal1 ;thicknesscorrespondinglytosolvetheheattransfer

    ;question.Pressureoutletboundaryconditionsare ;introducedinthegasandwateroutletplanes. ;DETERMINATIONOFPERFORMANCEPARA

    ;METERS

    ;Thepressuredropandheattransferperformance ;arecharacterizedbyafrictionfactorfandanaverage ;heattransfercoefficienth,respectively.Theflow ;conditionscanbecharacterizedbyReynoldsnumber. ;FollowingtheconventionslaiddownbyDonget ;a1.(2007),Renumberisestimatedasfo1lows: ;Re=pu</p,(5)

    ;whereuistheflowvelocityonshellside,deisthe ;hydraulicdiameteroftheshellside.Forthemodels ;usedinDongeta1.(2007),

    ;=

    ;2(R7r)/(R+7r),(6)

    ;whereRistheshe11outsideradiusandristhetube ;radius.

    ;Shellsidefrictionfactorfcanbewrittenas ;f=(LId

    ;~

    ;)(pu/2)(7)

    ;Heatabsorbedbywaterandsuppliedbyhotgas, ;QcandOhcanbewrittenas(EiamsaardandProm

    ;vonge,2007)

    ;Qc=Cp,.(rcrc),Oh=rhnCplh(,in),(8)

    ;1273

    ;whereandrhharethemassflowrateofwaterand ;gas,respectively;Cp,candCp,harethespecificheatof ;waterandgas,respectively.Theaverageheattransfer ;coefficienthisestimatedasfo1lows:

    ;where

    ;h=Q/[A(Tbwau)],

    ;=(.+jn)/2,

    ;O=(Qc+Qh)/2,

    ;(9)

    ;(1O)

    ;(11)

    ;wlistheaveragewalltemperatureattheouter ;surfacesofthetubeandbaffle,Aisthesummationof ;alltheheattransferareas(thehelicalbaffleandseven ;tubesareincluded).

    ;NIMERjCALRESULTS

    ;Inthepresentstudyfivecaseswithdifferent

    ;watermassflowratesaresimulated.Thesteadystate ;resultsarelistedinlb1e1.whichexhibitsthe ;volumetricwaterflow,thepressuredropofshellside ;andthetemperaturecontrast.

    ;Fig.5sketchesthepathlinesinbothwaterand ;gasflowfields.Thenon.arrowlinesfigurethepath ;lineofthegasflowintubesideconversetotheY-axis ;direction,whichisalmostthesameinthethree ;Table1Simulatedresultsofdifierenteases ;AP:pressuredifference;AT.-temperaturedifference ;

    ;1274Huangeta1./JZhejiangUnivSciA20089f9J:1270-1276 ;(a)(b)(c)

    ;Fig.5Pathlinesofvariousmodels(h=0.89m/h).(a)ModelA;(b)ModelB;(c)ModelC ;models.The~rrowlinesshowthestreamlinepattem ;ofthewater.Thesefiguresshowthatthewaterflow ;fieldsaregreatlyinflectedbyusingthehelicalbame. ;whichextendstheflowdistanceandstrengthensthe ;circleflowaccordingly.Theshellsideflowinmodel ;BismoredisorganizedthanthatinmodelC.which ;demonstratesthatthetangentialwaterinletdesign

    ;canadiusttheflowpattemanddecreasetheturbu

    ;lenceintensityeffectively.

    ;Fig6shoWsvelocityvectorsonX.zplane(ex

    ;actlyinthemiddleofcooler)undertheconditionof ;=

    ;1.5m/h.A11thevectorsaredisplayedtothe ;samescale.AsseeninFig.6a.theshortvectorsindi

    ;catethatmostofthevelocityvectorsareuprightto ;X-ZplaneinmodelA.InFigs.6band6c.theflows ;alongthehelicalbamesareaccompaniedbyrotation. ;secondaryflowsaregeneratedandtheflowrunspar

    ;alleltotheX-Zplane.Thetemperaturedistribution ;andpathlinesinpartofZsection(modelB)are ;showninFig.7.Thisfigureillustratesthegeneration ;ofavortex.Accordingtothefieldsynergyprinciple ;(Guoeta1.,l998).theoverallheatexchangeeffi

    ;ciencyofthecoolerisinfluencedbythesynergY ;betweenthefloworientationandthefieldtempera

    ;turegradients.AsshowninFig.7.theflowandtem

    ;peraturefieldsmatchfairlywel1.Basicallythetem

    ;peraturegradientsareverticaltothetubewallinthese ;cases.Afunctionofthehelicalbafneistocoordinate ;theflowandtemperaturefieldstosatisfythefield

    ;synergyprinciple.Italsoaccountsforthedistinct ;enhancementofheattransferefficiency. ;Thetemperatureandheattransfercoefficientin ;thehelicalbafneareobservedtobemuchlowerthan ;thatinthetubesurface(Fig.8,.Thesefiguressuggest ;thattheprimaryfunctionofthehelicalbameisnei

    ;thertostrengthentheheatexchangeinitselfnorto ;widenthediathermanoussurface,buttoenhancethe ;totalefficiencyofheatexchangebyimprovingthe ;flowpattem.Becauseoftheorientedeffectofthe ;helicalbame.theflowintensityisweakenedinsome ;particularregion,whichmakesthetemperaturedis

    ;tributionasymmetric.Misdistributionsofflowand ;temperaturecausedbythehelicalbafflecanhavea ;badeffectoncoolerperforlTlanceandreliabilit,,. ;ModifyingtheintakeangleandpositioniSanalter

    ;nativeapproachtodealingwiththisproblem.Any

    ;way.theheatexchangeofthewholecoolerincreased ;byusingthehelicalbame.

    ;0

    ;(a)(c)

    ;Fig.6VelocityvectorsinX-Zplane(h=1.5m/h). ;(a)ModelA;(b)ModelB;(c)ModelC

    ;Fig.7Temperaturedistributionandpathlinesof ;modelB(inpartofX-Zplane,h=1.5m/h)

    ;

    ;Huangeta1./dZhejiangUnivSciA20089f9J:1270-1276 ;(a)

    ;Fig.8Distributionofthetemperatureandthesurface ;heattransfercoefficient(h0.89m/h).(a)Tem

    ;perature,modelA;(b)Temperature,modelB;(c) ;Heattransfercoefficient,modelA;(d)Heattransfer ;coefficient,modelB

    ;Thefrictionalinf1uencesofthreemodelsare ;displayedinFigs.9and10.AstheVOlumetricflow ;increased.theshellsidepressuredropincreased. ;Properattentionisrequiredinthattheshellside ;pressuredroDofmodelCislessthanthatofmodelA ;inthesamecondition.Theprobablereasonisthatthe ;tangentia1WaterintakeofmodelCcouldavoiddirect ;impactonthetubewal1.Hencethehydraulicimpact ;pressureisobviouslyreduced.Asshowninlb1e1. ;thepressuredroDofmodelCisabout25%lessthan ;matofmodelB.

    ;Fig.11plotstheheattransfercoefficientVSwater ;Reynoldsnumber.ThehighertheReynoldsnumber, ;thestrongeristheturbulentflOWgenerated.Theheat ;transferrateisenhancedcorrespondingly.Inthis ;evaluationmodelBexhibitsthebestheattransfer ;performanceofthethreemodels.Underthesame ;waterReynoldsnumber,theheatansfercoefficient

    ;ofmode1BiSabout4%higherthanthatofmode1C. ;especiallywithabigReynoldsnumber.Inmode1A ;theheattransfercoefficientiSabout4.2%to10.28% ;lowercomparedtomodelB.Actuallythedifference ;betweenmodelAandmodelBoftemperature(Table ;11ismuchmorenotablethanthatoftheheattransfer ;coefflcients.Thereasonisthatthetotalareaofseven ;1275

    ;tubesandhelicalbaffleareconsideredastheheat ;transferarea(factor1inmodelsBandC,butonly

    surfaceareaisaccountedforinmodelA.As ;thetube

    ;showninEq.(9),thebiggerthefactorA,thelessthe ;heattransfercoefflcientaccordingly.whichlessens ;theheattransfercoefficientdifferencebetweenmod

    ;elsAandB.

    ;Volumetricwaterflow(m/h)

    ;Fig.9PressuredroponsheHside

    ;8l2l62O24

    ;WaterRenumber(×104)

    ;Fig.10F~cflonfactorvsRenumber

    ;WaterRenumber(×104)

    ;F.11SheHsideheattransfercoefficientvsRenumber ;.10pl?IaI10盘口JpaJns?oJd

    ;_I3J?蠹扫0I10它一?一一QII?

    ;

    ;1276

    ;C0NCLUS10N

    ;Huangetal/JZhejiangUnivSciA20089(9):1270-1276 ;ThepresentstudydealswithaCFDanalyrsisfor ;theflowstructuresandtemperaturedistributionsof ;theEGRcooler.Thenumeric

Report this document

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