divergence-detected-in-AMG-solver-pressure-correct

问题：8-31-rke-steady-4-0.8运行至11000迭代步是报错“divergencedetectedinAMGsolver:pressurecorrection”（补充：每个工况均遇到此类情况，原因尚未找到，暂时解决办法是将压力松弛因子从0.3改为0.1）；解答：ThiserrorisreportedwhenFluentisnotabletoconvergethesolution。Theresidualsarenotconverged。Thiserrormaycomeduetovariousreasonslikeimpropermesh，improperboundaryconditions，wrongmaterialandimpropersolutionsettings。Itisadvisabletorechecktheentirecasewhenoneencountersthiserror。以上老外对这个问题的回复，确实很多客户遇到过这样的问题，原因是多方面的，主要原因集中在网格质量差和边界条件的不合理。对于你的问题，如果之前都没有错误，到一定计算步才出现这样的问题，说明不是边条的问题，问题可能出在局部的网格上面，你将压力松弛因子调小，增加跌代次数，减小了误差，是解决问题的方法之一。如果想彻底回避这样的问题出现，建议在网格上还需要下点工夫！附：老外对类似问题的回复，你可以参考一下。Itrytostudytheturbulenceinsideaverticalcylinder.Inafirsttime,Itakethecaseofsteadyflow.Theairisintroducedaxiallyfrombelow,deflectedbymeansofasmallconicaldeflectorandthankstothegeometryofthechambertheairgoesupintothecylinder.Itaketheidealgaslaw,k-epsilonmodel,inletandoutletpressureforboundaryconditions.MyproblemisthatIhavedirectlythemessagedivergencedetectedinAMGsolver:kwhenIwanttoiterate.ItryalotofsolutionsIhavefindinthisforumwithoutanyresult.Couldanyonehelpme?ThanksinadvanceCheckyourhardware,especiallyRAMsockets.Ihaveexperience,thatthiscouldbeinitiatedbysomebadmemoryaddresssectors.Ihadasimilarproblem,butwithpressurecorrection.IfoundthatIhadsomehighlyskewedcells.OnceIcorrectedthehighlyskewedcellsbyadaptingtheiso-valuesofcellskewnessmycasebegantoiterate.Perhapsthismaybeyourproblem.Firstinitializethesolutionandthengotocontoursgridandselectequiangleskewandclickcomputetogetyouminandmaxs.Havingskewedcellsof0.9orhigherisn';tgood,asinmycase.HopethishelpsIhavechecktheskewedcellsbutitappearsthatinmycase,thisnumberislowerthan0.4Isittoohigh?Hi.Ihaveasimulationofasupersonic-valverunningbuttheprogramshowsmethismessage:divergencedetectedinAMGsolver:temperatureItriedtorisethelimitsfortemperatureandothersolutionlimitsbutnothinghashelpedsofarDoesanyonewhattodothanksTrytousemoreconservativeUnder-relaxationfactor.Ihadthesameproblemwithawinginatransonicflight.Accordingtomethesegregatedsolverisnotsuitablefortheconditionswithhighcompressibility.Letmeknowifyousucceedinusingthesegregatedandhow...Isuggestusingthecoupledsolver，Yourproblemwillvanish...Hopetoheargoodnewsfromyousoon.Lucathiswillbesomethingwrongwiththeeitherboundaryconditionorwithmodelselection,generallyifthingsarerightfluentjustconvergesfinewithoutanywarningorproblem.inmyopinionthesearethesignsthatyouaredoingsomethingwronginsettingupthecase,lookcloselyHi,Ihavethesameproblem,andcouldn'tsolveitbyusingthecoupledsolver.Isthereanyotherpossibleway.Iamsimulatingofficeroomproblem.Itseemsverysimplebutoneverystageitiscreatingproblemsforme.Iamgoingtotrytochangetheunderrelaxation.letsseewhatcomesup.Thankslongago,oneofmyfriendwasdoingCFDanalysisofakitchen,,usingcoupledsolver,andhefacedthesameproblem,forsometimehedidn’tgettheconvergedresults.FinallywhenIlookedathismesh.Itseemedtheproblemiswiththemeshnotwiththesolversettings.Isuggestedhimtomakemeshmorefiner,andviolaitgaveresultsinonerun,withoutanyproblem.somoralofthestorycheckyourmesh.Mymeshisquitefine,butIwilltrymorefinermesh.Thanksthenitshallconverge..anywayhaveyoutriedusingsegregatedsolver,howtheresultswiththatError:DivergencedetectedinAMGsolver?PostedBy:frankDate:Sun,16Jan2005,3:58p.m.Whatdoesmeanandhowdoifixit?Iamruningarosselandradiationmodelanditwontrunpast1iteration.Frankyouareusingthesegregatedsolver,soyoumustreducetheunderrelaxationfactorasyoucan.Youcouldhaveagridissueconcerninghighskewness.Thiscanbecheckedbygoingtocontoursgridcellequiangleskewandselectcomputeandchecktheminandmax.Youshouldbebelow0.9.Ifoverthenthiscouldbeyourproblem.Ithashappenedtomeacoupleoftimes.thanksamillioniwilltrythis...frankAMGsolver:kdivergence?PostedBy:RobertDate:Sun,9Mar2003,11:23a.m.HowdoIremedyadivergencewiththismessage.Thisisforscramjetcombustionin2Dwithinjectorsonboththetopandbottom.Isthereanywaytolookuperrormessages?reduceunderalaxtionfactor.Ihavetriedreducingtheunderelaxationfactorsbutthedivergencepersists.Anyothersuggestions?PerhapsintheAMGsolverMenu?Itlookslikeaproblemwithyourboundaryconditions,besurethey'recompatiblewitheachotherandconsistentwiththephysicsofwhatyou';retryingtosimulate,Ithinkyou'dbetterdon'ttouchanythingintheAMGmenu,unlessyouknowexactlywhatyou'redoing,Itismyopinion,ifthiscouldhelp.请求高手指点：循环流化床中非稳态气固两相流计算，使用欧拉双流体模型，K-E湍流模型，网格质量0.55。先计算了层流稳态，残差小于0.01时转为湍流非稳态。计算时总是出现divergencedetectedinAMGsolver:pressurecorrection。收敛因子已经调小，请求高手指点。万分感谢！如果你的网格质量没有问题（Theskewnessmustbelessthan0.85forQuad/Hexand0.9forTri/Tetcells），那就可能是压力速度耦合方式的问题，你也可以检查下你的压力插值格式是否适合多相流计算；也有可能是边界条件设置的问题。下面是论坛上的回答，我觉得讲的挺有条理的，这个是可压缩计算出现这个警告的解决回答，你看看对你有用的方面就行了。Well,thismessageshowsthattherunisdiverging,namelyinpressurecorrectionequation(Sec.25.4.3Pressure-VelocityCouplingofFluentUsersGuide).It'shardtotellwhythisishappeningwithoutcheckingyourcase.Areyoufollowingsomeofthesolutionconvergencestepsmentionedinthispage?-(1)Whatpressure-velocitycou