神經(jīng)網(wǎng)絡逆系統(tǒng)軟測量方法及其在紅霉素發(fā)酵過程中的應用.pdf

1、東南大學碩士學位論文神經(jīng)網(wǎng)絡逆系統(tǒng)軟測量方法及其在紅霉素發(fā)酵過程中的應用姓名：禹東川申請學位級別：碩士專業(yè)：控制理論與控制工程指導教師：戴先中20030701東南夫學碩士學位論文AbstractTheprimarycontroIvariablescannotmeasureoo】ineinrealbioprocesscoutro】applicationTodealwiththeproblem，anoveltheorySMNNIS(sofl

2、—measurementbasedonneuralnetworkinversesystem)，whichoriginatesfromtheadvanceresultofourlaboratoryisadvancedinmythesisandisappliedtotheerythromycinfedbatchfermentationprocessinaPhysicCoThegoodresultapprovestheprimarybiolo

3、giccontrolvariables，includingmycetiumconcentration，sugarconcentrationandproductionconcentrationCallbeestimatedonlinebytheSMNNISSotheSMNNIScanbea5anewmethodtodealwiththeestimateofbiologiccontrolvariableinsimilarbioprocess

4、ThemainresearchfindingsinmythesisIistasfoflowsIThattheSMNNIS(SoftMeasurementbasedonNeuralNetworkInverseSystem)originatesfromtheadvanceresultofourlaboratoryisadvancedThegoodresultapprovestheprimarybiologiccontrolvariables

5、，includingmyceliumconcentration，sugarconcentrationandproductionconcentration，canbeestimatedonlinebytheSMNNISSotheSMNNIScanbeasanewmethodtodearwiththeestimateofbiologiccontrolvariableinsimilarbioprocess2Asuitablenon—struc

6、turekineticsmodelingoferythromycinfedbatchfermentationprocessisbroughtforwardbyreadingadvancereferenceandtherealmanufacturingtechniqueandispreparedforthesoft—sensorandoptimizingcontrolinthenearfuture3Thestructuredecompos

7、itionof／leuralnetworkisintroducedtoimprovetheeonvergencerateandgeneralizationabilityofneuralnetworkbecausetheneuralnetworkintheSMNNISisusedforapproximationofMIMOfunction4TheprogramoftheSMNNISbeingusedfortheerythromycinfe

8、d—batchfermenmtionprocessisdesignedbaseonthemodularizationtheoryThearithmeticoftheSMNNIScanbeasastandardmoduletobetransplantedtosimilarproblembyalteringjustalittleKeywords：artificialneuralnetwork，inversesystem，softsensor