Title of article
Analysisofthecriticalheatfluxinroundverticaltubesunderlowpressureandflowoscillationconditions.ApplicationsofartificialneuralnetworkReviewArticle
Author/Authors
SuGuanghui، نويسنده , , KMorita، نويسنده , , KFukuda، نويسنده , , MarkPidduck، نويسنده , , JiaDounan، نويسنده , , JaakkoMiettinen، نويسنده ,
Issue Information
روزنامه با شماره پیاپی سال 2003
Pages
19
From page
17
To page
35
Abstract
Artificialneuralnetworks(ANNs)forpredictingcriticalheatflux(CHF)underlowpressureandoscillationconditionshavebeentrainedsuccessfullyforeithernaturalcirculationorforcedcirculation(FC)inthepresentstudy.TheinputparametersoftheANNarepressure,meanmassflowrate,relativeamplitude,inletsubcooling,oscillationperiodandtheratiooftheheatedlengthtothediameterofthetube,L/D.TheoutputisanondimensionalizedfactorF,whichexpressestherelativeCHFunderoscillationconditions.BasedonthetrainedANN,theinfluencesofprincipalparametersonFforFCwereanalyzed.TheparametrictrendsoftheCHFunderoscillationobtainedbythetrainedANNareasfollows:theeffectsofpressurebelow500kPaarecomplexduetotheinfluenceofotherparameters.Fwillincreasewithincreasingmeanmassflowrateunderanyconditions,andwilldecreasegenerallywithanincreaseinrelativeamplitude.Fwilldecreaseinitiallyandthenincreasewithincreasinginletsubcooling.TheinfluencecurvesofmeanmassflowrateonFwillbealmostthesamewhentheperiodisshorterthan5.0sorlongerthan15s.TheinfluenceofL/DwillbenegligibleifL/D>200.Itisfoundthattheminimumnumberofneuronsinthehiddenlayerisaproductofthenumberofneuronsintheinputlayerandintheoutputlayer.
Journal title
Nuclear Engineering and Design Eslah
Serial Year
2003
Journal title
Nuclear Engineering and Design Eslah
Record number
894494
Link To Document