《sensor Based on Neural Networks in Rare-earth Countercurrent Extraction Process.doc》由会员分享,可在线阅读,更多相关《sensor Based on Neural Networks in Rare-earth Countercurrent Extraction Process.doc(16页珍藏版)》请在金锄头文库上搜索。
1、sensor Based on Neural Networks in Rare-earth Countercurrent Extraction ProcessVol_32.No.4ACTAAUT0MATICASINICAJuly,2006ComponentContentSoft-sensorBasedonNeuralNetworksinRare.earthCountercurrentExtractionProcess1)YANGHuiICHAITianYlOu(SchoolofElectronicsandElectricalEngineering,EastChinaJiaotongUniver
2、sity,Nanchang330013)(ResearchCenterofAutomation,NortheasternUniversity,Shenyang110004)(E-mail:)AbstractThroughtfusionofthemechanismmodelingandtheneuralnetworksmodeling.acompo-nentcontentsoftsensor.whichiscomposedoftheequilibriumcalculationmodelformulti-componentrareearthextractionandtheerrorcompensa
3、tionmodeloffuzzysystem,isproposedtosolvetheprob-lemthatthecomponentcontentincountercurrentrare-earthextractionprocessishardlymeasuredon-line.AnindustryexperimentintheextractionyprocessbyHABusingthishybridsoft-sensorprovesitseffectiveness.KeywordsRare-earth,countercurrentextraction,softsensor,equilib
4、riumcalculationmodel,neuralnetworks1IntroductionChinahasthemostabundantrareearthresourceintheworldl.Buttheextractionprocessautomationisstillinthestagethatcomponentcontentismeasuredoff-line,andtheprocessiscontrolledbyexperienceandparametersareregulatedbyhands.Thissituationleadstolowefficientproductio
5、nrate,highresourceconsumptionandunstableproductionquality,andlimitstherareearthindustrydevelopmentt1.T0implementautomationintherareearthextractionprocess.theonlinecomponentcontentmeasuringmustbeachievedatfirst.MaincomponentcontentmeasuringmethodsincludetheXrayabsorbspectrometer,Xrayenergyspectromete
6、r,IsotopicX-fluorescenceenergyspectrometer,on-linespectrophotometers.etc,.Theseinstrumentshavenotbeenwidelyusedinindustrybecauseofhighcost,lowreliabilityincontinuousoperation,complexmaintenance,delayinmeasurementandlowaccuracy.Thesoftsensormethod,ontheotherhand,providesanewwaytoon-linemeasurecompone
7、ntcontentintherare-earthcountercurrentextractionprocess.Wlewillfurtherourresearchin4byfusionofthemechanismmodelingandtheintelligentmodelingandproposeahybridsoftsensoroftherareearthcomponentcontentwhichcontributestobetterpredictionaccuracyandwiderapplicability.AnindustryexperimentintheextractionYproc
8、essbyHABusingthishybridsoftsensorprovesitseffectiveness.2Descriptionofrare-earthextractionprocessAtwocomponent(AandB)countercurrentextractionprocessisshowninFig.1,whereAistheeasilyextractedcomponentandBisthehardextractedcomponent.Theleftsideistheextractionsectioncomposedof礼stagemixclarifiers.Therigh
9、tsideisthescrubsectioncomposedofmstagemixclarifiers.InFig.1,u1istheflowofrareearthfeed,u2istheflowofextractionsolvent,u3istheflowofscrubsolvent,u4andu5arethedistributionsofAandBinthefeed,respectively,whereu4+u5=1.pAistheorganicphaseproductpurityofAattheexitandPBistheaqueousphaseproductpurityofBatthe
10、exit.PA,kisorganicphasecomponentcontentatthespecifiedsamplingpointinscrubsectionandpB,isaqueousphasecomponentcontentatthespecifiedsamplingpointinextractionsection.Sincethewholeprocessiscomposedofdozensofuptoonehundredstages,theflowregulationofextractionsolvent,scrubsolventandthefeedcannotinfluenceth
11、eproductpurityattheexituntilalong-timedelay(oftenhoursevendays).Fortheabovereason,thesamplingpointissetneartheexitandtheexitproductpurity(pA,pB)isguaranteedbymeasuringandcontrolofthecomponentcontents(pA,k,pB,k)atthesamplingpoint.Accordingtocountercurrentextractionprinciple,theparametersp(pA,korPu,k)
12、,Ul,u2,u3andu4havethefollowingnonlinearrelationshipP=f(ui,u2,u3,u4,1)Supportedby丽onalNaturalScienceFoundationofP.RChina(50474020,60534010,60504006)ReceivedJanuary24,2005;inrevisedformFebruary16,2006(1)490ACTAAUToMATICASINICAVo1.32where,_isthecomplicatednonlinearfunctionandisthedisturbance.Howtomeasu
13、retheparametersp(pA,korPB,)hasbecomethekeypointofrare-earthprocessautomation.Fig.1Rareearthcountercurrentextractionprocess3MethodofcomponentcontentSOft-sensorintherareearthextractionprocess3.1StrategyofcomponentcontentSOft-sensorinrareearthextractionprocessSincethecountercurrentrareearthextractionpr
14、ocesshasnonlineardynamicsanduncertainty,itisdifficulttomode1thisprocessbysimplemethodsincludingmechanismmodelingorparametersestimationmethodstTheframeworkofrare-earthextractioncomponentcontentSOft-sensorsystemisdescribedinFig.2.Fig.2FrameworkofrareearthextractioncomponentcontentsoftsensorThewholesys
15、temiscomposedofdatasamplingandpretreatmentsubsystem,extractionequilibriumcalculationmode1anderrorcompensationmode1usingadaptiveneuralfuzzynetworks.ParameteriSthecomponentcontentassayvalue,parameter芦iStheoutputofcountercurrentextractioneqm一1ibriumcalculationmodel,:P一isthemodelingerror,andparameteristheoutputofcomponentcontenterrorcompensationmode1.Apd=一iSusedtotrainthemodelingerrorcompensationmode1.ThentheSOftsensoroutputp=卢+(2)3.2Equilibriumcalculationmodelformulticomponentcountercurrentrare-earthextractionFortherare-earthcountercnrrentextractionprocessshowninFig.1,whenthecomponentAandBrea