基于灰色區(qū)間數(shù)學(xué)模型的電站設(shè)備機(jī)況評估.pdf

1、上海交通大學(xué)博士學(xué)位論文基于灰色區(qū)間數(shù)學(xué)模型的電站設(shè)備機(jī)況評估姓名：周宇陽申請學(xué)位級別：博士專業(yè)：動(dòng)力機(jī)械與工程指導(dǎo)教師：陳漢平20020401rerformanceestimationofmetaIapparatusorthermodynamicsysteminpowerplantswillinevitablemeetuntortaintiesForexampletheheattransferportialdifferentialeq

2、uationreliedontocalculateandestimaterotor’stemperaturefieldsandconsequentIhennalStItSSeSalwayshasitsboundaryandinitialconditionsgiveninimprecisetermsduetothedi幣cultiesofachievinganaccuratecharacterizationofthosefacilitie

3、splayingimportantroleslnoperation，maintenanceisalwaysimplementedbasedonengineersjudgmentsandexperiences，etcThoseuncertaininputsarecommonlyreplacedbycrispnumberstosimplifythecalculationprocedurebutinthecostofunknownprecis

4、ionofIheoutputsInthispaperwetrytodescribethoseuncertaintiesasintervalnumbersbyIntetvaIArithmeticbasedGrayModel(IAGM)，consequentlyconstructandsolvethethermalstressesmodelwithintervaIarithmetic”11irstthemathematicaIanalyti

5、cresultofatypicalnonborerotorisdeducedtoshowthemonotoniccharacteristicsofsimplegrayboundaryinfluencedIAGM，thusanEnd—PointMethod，whichrepresentstwoendsofeachoutputindividuallybycoincidedendsoftheinputs，canbeappliedAccordi

6、ngtothecomplexitiesofcompoundgraymetalpropertiesandgrayboundariesinfluencedIAGMcorrespondingfinitedifferenceequationsinCNformatarecreatedThenOettli‘sviewpointsareextendedtoanImplicit—InequalitiesbasedMethodinordertogetth

7、esolutionrapidlyandeffieientlyNumaricexamplesofSTC600MWandMitsubishi700MWsteamturbinerotorswhichisconditionedtosteppedsteamtemperatureonboundariesarebroughtintoillustrationr111hefiniteelementmethod，withitsfiexibilityinde

8、alingwithcomplexgeometries，isfrequentlyutilizedintemperaturefieldscomputingStimulatedbytheSuccessfulcombinationofjntervalarithmeticandfinitedifferencemodel，agrayfiniteelementmethod(gFEM)modeIisconstructedfortransientther

9、maIanalysisConsideringtheunavoidableintervaIexpandingduetointrinsiccharacteristicsofjntervalarithmetic，especiallywheniteratinglargescalesparselinearequations，weconvertthemodeltoaminimaxprogrammingintheoptimalsenseAtwicea

10、ppliedLimethodwhichtriestosubstitutetheminimaxprogrammingtoanunconstrainedoptimizationisfoundmoree塌cientthanthegeneralsequentialquadraticprogrammingTheresultsinfoimsofintervalnumbersshouldbemorerepresentativeforthewholes

11、pectrumofpossibleonesBynumeralexamplesshownitisbelievedthatthegFEMmodelisapplicableinpracticeifthecomputingtimeisconfirmedtobemoreacceptableIL廳lVloreparameters，whichmostlyobtainedfromIpealmeasurement，areinvolvednthermale

12、conomicaIemciencyanalysismodeIconstruction、Simultaneouslythosededucedparameterssuchascnthalpyandfluxofkeystream，exeganimportantinfiuenceonthefinalresultsApplicationofintervalarithmeticproposedjnthispaperdealsthemodelwith

13、sequentiaIflowsheetingthusderivationof“intervalenthalpyandfluxcanbemadeAtierbeingtreatedinoptimizationsensetheheatrateiseasilYcalculatedComparisonoftheresultsgainedrespectivelyintraditionalandlAGMwayshowstbeobvioussuperi

14、oritytoerrortractableanderrormanageableofthelaReroner111hepreviouslyapplied01logichasmadetheforinerfaultdiagnosisalwaysresultinwrongoutcomewhichpreventexistingmodelsfronlwidelypracticalapplicationsIAGMjsfoundsuitforinput

15、parametersandcomputingarithmeticbetailoredtoowetheabilityofdescribingthevaguenessanduncertaintytheinformationwilIbeutilizedadequatelytomakethemodeImorebeneficialIntervaImllnbersreplacesthe01IogicaIinputswhilethearithmeti

16、cisamendedtograydistanceiudgmentcriteriaInanotherwaytheformerneuralnetworkbasedfaultdiagnosismodelhasbeenimprovedona業(yè)rayintervalneuralnetworkWiththesenewmodelspresentedinthispapermisident訊cationcausedbyvagtieinputsandres

17、emblingmembershipwouldbeeliminatedtoalowerIeveIthusmakethediagnosismoreprecise關(guān)鍵詞Keywords扶色理論區(qū)間代數(shù)數(shù)學(xué)模型電站機(jī)況評估熱分析性能計(jì)算故障診斷GrayTheoryIntervalArithmeticMathematicModelPowerStationPerforInanceEstimationThermalAnalysisThermalEco