2001 analytical and neural network models for gas turbine design and off design simulation

《2001 analytical and neural network models for gas turbine design and off design simulation》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、Int.J.AppliedThermodynamics,ISSN1301-9724Vol.4,(No.4),pp.173-182,December-2001AnalyticalandNeuralNetworkModelsforGasTurbineDesignand#Off-DesignSimulation*AndreaLAZZARETTOandAndreaTOFFOLODepartmentofMechanicalEngineering-UniversityofPadovaViaVenezia1-35

2、131Padova–ItalyTel:+0390498276747,Fax:+0390498276785E-mail:andrea.lazzaretto@unipd.itAbstractThispaperpresentsagasturbinedesignandoff-designmodelinwhichthedifficultiesduetolackofknowledgeaboutstage-by-stageperformanceareovercomebyconstructingartificial

3、machinemapsthroughappropriatescalingtechniquesappliedtogeneralizedmapstakenfromtheliteratureandvalidatingthemwithtestmeasurementdatafromrealplants.Inparticular,off-designperformanceisobtainedthroughcompressormapmodificationsaccordingtovariableinletguid

4、evaneclosure.Thesetofequationsofthedevelopedanalyticalmodelissolvedbyacommercialpackage,whichprovidesgreatflexibilityinthechoiceofindependentvariablesoftheoverallsystem.Theresultsobtainedfromthissimulatorareusedforneuralnetworktraining:problemsassociat

5、edwiththeconstructionanduseofneuralnetworksarediscussedandtheircapabilityasatoolforpredictingmachineperformanceisanalyzed.Keywords:analyticalmodel,neuralnetwork,designandoff-designperformance,gasturbine1.IntroductionThegasturbinedesignandoff-designmode

6、lpresentedinthispaperaimsbothatcomputationalTheincreasingsuccessofgasturbinesinsimplicityandattheabilitytodealwithplantsindustrialapplications(bothelectricandthermalhavinglargevariationsintheoperatingparameters.energyproduction)ispartlyduetotheirquickT

7、hemainissueinthedevelopmentofanoff-designresponsetoloadvariations;infact,theseplantsaremodelisthepredictionofmachineoff-designoftenusedforelectricproductioninanintermediateperformance.Themodelshouldbeconstructedrangebetweenpeakandbasedemand.usingthesta

8、ge-stackingtechnique,simulatingtheMostgasturbineanalyticalmodelsproposedperformancemapsofeachmachinestage,andthenintheliteraturearesimplydesignmodels.Someverifiedusingoperationaldataonsingleothers,however,areabletopredictoff-designcompr