資源描述:
《基于逆系統(tǒng)的感應(yīng)電機(jī)神經(jīng)網(wǎng)絡(luò)解耦控制》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、東北大學(xué)碩士學(xué)位論文摘要基于逆系統(tǒng)的感應(yīng)電機(jī)神經(jīng)網(wǎng)絡(luò)解耦控制摘要三相交流感應(yīng)電機(jī)由于其可靠、堅(jiān)固而且成本相對(duì)較低的特性而廣泛應(yīng)用于工業(yè)生產(chǎn)各個(gè)領(lǐng)域。由于它是一個(gè)多變量、強(qiáng)耦合、非線性系統(tǒng),當(dāng)運(yùn)行過(guò)程中某些參數(shù)變化時(shí),采用常規(guī)控制方式不能及時(shí)調(diào)整控制參數(shù),無(wú)法滿足高性能的調(diào)速要求。因此設(shè)計(jì)出高精度、適應(yīng)能力強(qiáng)的感應(yīng)電機(jī)控制系統(tǒng)的要求越來(lái)越迫切了。本文首先使用逆系統(tǒng)方法對(duì)感應(yīng)電機(jī)調(diào)速系統(tǒng)的數(shù)學(xué)模型進(jìn)行可逆性分析,在理論分析的基礎(chǔ)上,使用BP靜態(tài)神經(jīng)網(wǎng)絡(luò)加積分器的方法來(lái)構(gòu)造原系統(tǒng)的口階逆系統(tǒng),利用壓縮映射遺傳算法優(yōu)化神經(jīng)網(wǎng)絡(luò)的初始權(quán)值和閾值,給出了構(gòu)造神經(jīng)網(wǎng)絡(luò)a階逆系統(tǒng)的具體的方
2、法、步驟、設(shè)計(jì)原則和注意事項(xiàng)。再把得到的逆系統(tǒng)與原系統(tǒng)復(fù)合,將系統(tǒng)線性化解耦為轉(zhuǎn)速與轉(zhuǎn)子磁鏈兩個(gè)相對(duì)獨(dú)立的偽線性子系統(tǒng)。最后分別設(shè)計(jì)線性閉環(huán)調(diào)節(jié)器對(duì)解耦后的兩個(gè)子系統(tǒng)進(jìn)行控制。仿真是在MATLAB中進(jìn)行的。通過(guò)對(duì)感應(yīng)電機(jī)在額定參數(shù)和負(fù)載變化條件下兩組實(shí)驗(yàn)的比較,表明該設(shè)計(jì)使用的神經(jīng)網(wǎng)絡(luò)a階逆系統(tǒng)方法較好地實(shí)現(xiàn)了轉(zhuǎn)速與轉(zhuǎn)子磁鏈間的動(dòng)態(tài)解耦,系統(tǒng)對(duì)負(fù)載的擾動(dòng)有較強(qiáng)的抑制作用,系統(tǒng)的動(dòng)靜態(tài)特性明顯改善。實(shí)驗(yàn)結(jié)果證明,本文提出的神經(jīng)網(wǎng)絡(luò)口階逆系統(tǒng)方法具有較好的應(yīng)用前景。關(guān)鍵詞:神經(jīng)網(wǎng)絡(luò);逆系統(tǒng);感應(yīng)電機(jī);轉(zhuǎn)速;轉(zhuǎn)子磁鏈;解耦控制一II—NeuralNetworkDecoupling
3、ControlofInductionMotorBasedonInVerseSystemAbstractThfee.phasealtematingcullrentinductionmotoriswidelyusedinindustry'becauseOfitsfeliability,mggedization,andthecostofitisrelatiVely10w.It’samultivariabIe,stmn舀ycoupled,nonlinearsystem.Whensomeparameterschange,theconVentionalcontrolmethodcan’
4、tadjustthecontrolparameterintimeduringthemotornlnning.Itcan’tsatisfythehi曲requirementofspeedregulation.Sotodesigntheinductionmotorcontrolsystemwithhigllprecisionandstrongadaptabilityismoreandmorcu唱ent.First,analyzethereVersibilityofthemathematicalmodelofinductiOnmotorspeedregulatingsystem,
5、withtheinVersesystemmethod.Onthebasisoftheoreticanalysis,fabricatethea一歷orderneuralnetworkinversesystemoftheoriginalsystemwiththestaticneuralnetworkOfBPandinte舀atOrsconstmcted,andoptimizetheweiglltandthethresholdofneuralnetworkwithContractiVemappinggeneticalgorithms.Thenthemethods,thesteps
6、,thedesignprinciples柚dthecautionsofhowtoconstnlctthea一咖orderneuralnetworkinversesystema陀酉Ven.CascadingtheneumlnetworkiIlversesystemwiththeori酉nalsystem,thesystemisdecoupledintot、ⅣoindependentpseudO—linearsubsystems,speedsubsystem柚drotornuxsubsystem·Thenalinearclose—loopadjustorisdesi印edtoc
7、ontroleachofthedecoupledsubsystemS.T1lesimulationisdoneiIlM:ATLAB.Comparedthet、vogmupsofexperimentsofinductionmotofinratedpar鋤etersandinloadvariation,thisdesj印methodhaspreferablyimplementedthedynamicdecouplingbetweenthespeedandrotornux,withthemethodoftheneural