预览加载中，请您耐心等待几秒...

基于模糊神经网络α阶逆系统的发酵过程多变量解耦控制(英文).docx

预览
1/2
2/2

在线预览结束，喜欢就下载吧，查找使用更方便

下载提示 文本预览

如果您无法下载资料，请参考说明：

1、部分资料下载需要金币，请确保您的账户上有足够的金币

2、已购买过的文档，再次下载不重复扣费

3、资料包下载后请先用软件解压，在使用对应软件打开

基于模糊神经网络α阶逆系统的发酵过程多变量解耦控制(英文) Fermentationisanimportantprocessinthefoodandpharmaceuticalindustries,whichinvolvestheconversionoforganicsubstancesintoothercompoundsusingmicroorganisms.ThereactionkineticsandyieldoftheprocessdependonvariousfactorssuchaspH,temperature,andnutrientconcentration.Multivariablecontroliscriticaltoachieveoptimalprocessperformanceandproductquality. Fuzzyneuralnetworkshavebeenwidelyusedinprocesscontrolduetotheirabilitytohandleuncertaintyandnonlinearrelationships.Inthispaper,weproposeafuzzyneuralnetworkα-orderinversesystemformultivariabledecouplingcontroloffermentationprocesses.Theproposedapproachcaneffectivelycompensateforthecross-couplingeffectsamongdifferentcontrolloops,andimprovethecontrolperformance. Thefuzzyneuralnetworkα-orderinversesystemconsistsoftwoparts:theforwardmodelandtheinversemodel.Theforwardmodelisusedtoestimatethestateofthefermentationprocessbasedontheinputandoutputvariables.Theinversemodelisdesignedtogeneratethecontrolinputsignalsbasedonthedesiredsetpointsandtheestimatedstatevariables.Theinputandoutputvariablesarepreprocessedusingfuzzylogictohandletheuncertaintyandnonlinearityoftheprocess. Theproposedapproachwastestedonasimulatedfermentationprocesswithtwoinputs(pHandtemperature)andtwooutputs(biomassandproductconcentration).Thesimulationresultsshowedthatthefuzzyneuralnetworkα-orderinversesystemcouldsuccessfullydecouplethecontrolloopsandachievesatisfactorycontrolperformance.ThetrackingerrorsofthecontrolledvariablesweresignificantlyreducedcomparedtotheconventionalPIDcontrolmethod.Theproposedapproachalsoshowedbetterrobustnesstoparametervariationsanddisturbancerejectioncapability. Inconclusion,thefuzzyneuralnetworkα-orderinversesystemisapromisingmethodformultivariabledecouplingcontroloffermentationprocesses.Futureworkcanfurtherinvestigatetheapplicabilityoftheproposedapproachinreal-worldindustrialfermentationprocessesandexplorethepotentialofintegratingotheradvancedcontroltechniques.