基于支持向量機的稀土萃取過程組分含量軟測量研究

1、華東交通大學碩士學位論文基于支持向量機的稀土萃取過程組分含量軟測量研究姓名：陸榮秀申請學位級別：碩士專業(yè)：交通信息工程及控制指導教師：楊輝20070609SOFrSENSORMETHODOFCOMPONENTCONTENTINRAREEARTHSEPARArIONPROCESSBASEDONSVMABSTRACTTherareearthseparationprocessbycountercurrcnta【打ac6∞ischaracter

2、izedofnonemptytimevariantpropertiesandseverelagIntheprocessofmonitoringandcontrollingtherareearthsepalationbycountercurrentextractiontheonlineinformationofcomponentcontentisinaccessibleAtpresenttheonlineexaminationmethod

3、sofrealizingthemre圮arthseparationprocessbycountcrcmTentextractionhavetheshortcomingsofhighinstallmentpri=andcomplexsuuctureandlowreliabilityofthesystemcontinuousrunninganddi伍celtymaintenanceandSOOILBecauseithasprecisionr

4、eliablationeconomyandtherapiddynamicresponseandothercharacterssoftsensormcghodturnsintoanewapproachtealizetheOnlineestimationofcomponentcontentButmodelingofsoftsensoristhecoreofsoftSensorrea[izafioiLThekeytealizesoftsens

5、oristobuildthesoftsensormodelofcomponentoontenLButSupportv“l(fā)ofm硝血in鶴(icSVM)istoseekbestco=promjscaccordingtothelimitedsampleinformationinbetweenthemodelcomplexityandlearningcapabilityobtainthebestgeneralizationabilityIth

6、asalreadyusedinthenonlinearsystemmodelingwidelyIt缸mappingtheactualproblemintoahighdimensionalfeaturespacethroughthenonlineartransformationinwhichanoptimalseparatinghyperplaneOr矗lncli∞r(nóng)egressionisdoneItscomplexityoftheari

7、thmetichasnothingtodowiththeSampledimensiOnItmayob(amovelallopli力aa]solutionandsolvethepartialminimmproblemwhichisunabletoavoidintheneuralnetworkmethodInorderto∞qlIiIctherealtimedataandoontrlbutetealtimedosedloopcontrolt

8、histhesisdevelopmentsthesoflseasorandtheresearchofapplicationandsimulationtoSVMinthe栩礬卜canhseparationbycountercurrentextractionTheprimarycontentsareasfollows：1AfternarratingthepIoc搬examinationandcontrolactualityofrareear

9、thseparationitproposeswiththeideaofestablishingthesoftsensormodelbased∞SVMinviewoftheproblemofrealizingtheoldineestimationofcomponentcontentandreviewstheprincipleandtheresearchactualityofSvMtechnologyZItdepictstheprocess

10、ofrxreeanhseparationthenanalysisesthefactorswhichinfluencethecomponentcontentthroughmedhanj岱analysisoftheprocessofrareeachseparatiom11地softS即lSOrmodelbasedonSVMisimplementedtorealizeonlineestimationofcomponentcontentinra

11、reearthseparationprocess3Withthemodelsshortcomingoftheslowoperatingsp。cdandinfirmexudesabilj吼threeimprovedmethodstObuildsoftsensormodelofsvM(=ix=ikernel，LS_SVM，incrementalleaning)isinvestigatedandv盯i丘cdinsimulationThcsim