基于單神經元PID的永磁同步電動機速度控制器設計.pdf

1、分類號：TP273密級：——單位代碼：]魚359學號：—0404042—5100143套肥戈母火警HefeiUniversityofTechnology碩士學位論文MASTERDISSERTATIoNo論文]題目：基于單神經元PID的永磁同步電動機速度控制器設計學位類別：學歷碩士學科專業(yè)：(工程領域)鱉創(chuàng)堡迨生蕉魚蜓塑作者姓名：陳曼導師姓名：蔓鎖塞副教授完成時間：壘塑墨曼旦——PermanentMagnetSynchronousMoto

2、rDriveWithSingleNenronPIDSpeedControllerAbstractRecently，withtheadvancesinpermanentmagneticmaterials，semiconductorpowerdevicesmanufacturingtechniquesandcontroltheories，thepermanentmagnetsynchronousmotor(PMSM)driveobtaini

3、ngpopularityinmotioncontrolapplicationsPMSMsarewidelyusedinhighperformanceservoapplicationsbecauseofitscompactsize，highairgapfluxdensityandhightorque／inertiaratioBut，somemachineparameters(forexamplerotorresistance)anddra

4、gloadparameters(forexamplemomentofinertia)alwayschangewiththedifferentworkconditionsAllthecontroltheoriesbasedontheclassicalcontroltheoryandmoderncontroltheoryfacethesameproblems，inwhichtheyalldependonthemotormodel，andwh

5、enthesystemsufferedtheinfluencesofparametersvaryinganddisturbance，itsperformancewillbemuchaffectedSo，itisaninevitabletrendthatintroducesintelligentcontroltheoryintoACmotorcontrol，becauseitcangetridofthedependenceofcontro

6、lplants’modelandhasfeasibilityandrobustnessforprocessingtheuncertaintyandinexactproblemsPIDcontroliswidelyadoptedinindustrialfieldsbecauseofitssimplestructure，highreliabilityandeasilyimplementationConventionalPIDcontroll

7、erhasgoodcontroleffectiftheparametersofsystemmodelhavenotbigvariation，butwhencontrolledplanthashighnonlineartraitandhighuncertainty，PIDcontrolcan’tachievegoodresultNeuralnetwork，asoneofmoderninformationprocesstechnologie

8、s，hasvariousadvantagesinmanyapplicationsNeuralnetworkcontrolbecamearegardedresearchdirectionincontroltheorySingleneuronPIDControltheorycombinestheadvantagesofConventionalPIDcontrolwithneuralnetworkcontrolandhasthecharact

9、eristicsofsimpleconfiguration，easilyregulation，parametercanselfstudyandselfadaptInthispaperbasedonanalyzingofneuralnetwork，singleneuronPIDcontrolanditsireprovedstructureareintroducedAndbasedonthemodelingofPMSM，thesinglen

10、euronPIDspeedcontrollerwithadvancedHebbsupervisingstudyandsingleneuronspeedcontrollerbased011RBFneuralnetworkidentificationarestudiedandsimulatedbymatlabTheresultsdemonstratethatthesingleneuronPIDspeedcontrollerisimprove