氧化鋁生產(chǎn)四效逆流降膜式蒸發(fā)過程出口濃度預(yù)測模型研究

1、中南大學(xué)碩士學(xué)位論文氧化鋁生產(chǎn)四效逆流降膜式蒸發(fā)過程出口濃度預(yù)測模型研究姓名：畢慶華申請學(xué)位級別：碩士專業(yè)：控制科學(xué)與工程指導(dǎo)教師：唐朝暉20080528ABSTRACTThepracticesofproductionindicatedthat，intheevaporationprocessofaluminaproductionthefeedoutletconcentrationiSallimportant‘indicator；itiS

2、alsotheprimarybaseforregulatingtheoperationalparameters(flowofimportedfeed，thefreshsteamflowandpressure)Becausethemanualmeasurementofthefc；edoutletconcentrationlagbehindproductionprocessseveralhourslaterit’SdifficultforU

3、StomodifythemodelbyfeedbacktimelyFurtherowingtotheartificialregulationtosteamflowandotheroperatingparameters，theadjustmentofoperatingparametersforthecontrolofoutletconcentrationisalsobuiltontheexperienceoftheoperatorsand

4、theadjustmenttimeofsystemIStoolongInordertosolvetheproblemoftheoutletconcentrationforecastmtheevaporationprocessandestablishtherelationshipbetweenoutletconcentrationandtheoperatingparameters，theforecastmodeloftheoutletco

5、ncentrationoftheevaporationprocessisestablishedbystudyingthefoureffectbackfeedfallingfilmevaporationsystemofacertainaluminaplantT11iSpaperestablishesamechanismmodelbasedonmaterielbalance，heatbalanceandphasebalancebyanaly

6、zingthemechanismFurtherbasedonindustrialrunningdataaBPneuralnetworkmodelfortheforecastoftheoutletconcentrationisestablishedwhoseinputsarethedirectinfluencingfactorsoftheoutletconcentrationandtheoutputistheoutletconcentra

7、tiontooThecharactersofthetwomodelsareanalyzedrespectivelyandsomeconclusionsCanbedrawnasfollows：theBPneuralnetworkmodel’sfittingperformanceisfinebyandlargebutitcannotpredictexactlywhentheworkconditionsareunstablebecauseof

8、theshortageofdata；themechanismmodel’SprecisionislOWerthantheformergenerallybutitseffectiSbetterthanitwhentheworkconditionschangeabruptlySoallintelligentintegratedforecastmodeliSestablishedbycombinationthetwomodelsaboveAc