資源描述：
《基于改進(jìn)pso算法的過(guò)熱汽溫神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)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、基于改進(jìn)PSO算法的過(guò)熱汽溫神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)控制摘要火電廠鍋爐高溫過(guò)熱器的過(guò)熱汽溫是一類(lèi)非線性、時(shí)變性、大時(shí)滯、大慣性的典型對(duì)象。目前主要采用常規(guī)串級(jí)式PID控制，在干擾量大、工況發(fā)生變化時(shí)，難以達(dá)到理想的控制效果。神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)控制(NNPC)充分利用了神經(jīng)網(wǎng)絡(luò)的非線性映射能力以及預(yù)測(cè)控制滾動(dòng)優(yōu)化、反饋校正的機(jī)理，更符合此類(lèi)復(fù)雜系統(tǒng)控制的不確定性和時(shí)變性的實(shí)際情況。為了進(jìn)一步改善NNPc的性能，本文將具有全局搜索能力、實(shí)用性強(qiáng)的改進(jìn)粒子群優(yōu)化算法(MPSO)融合到NNPC中，提出了基于MPSO算法的RBF神經(jīng)網(wǎng)絡(luò)混合優(yōu)化(MPSO-RBF)
2、策略，構(gòu)造了基于MPSO．RBF混合優(yōu)化策略的模型預(yù)測(cè)器，以及基于MPSO算法的非線性優(yōu)化控制器。針對(duì)過(guò)熱汽溫的控制，給出了基于神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)控制的串級(jí)控制系統(tǒng)，并就該系統(tǒng)在實(shí)現(xiàn)時(shí)所涉及到的預(yù)測(cè)模型、滾動(dòng)優(yōu)化算法、反饋校正、仿真參數(shù)設(shè)置問(wèn)題等進(jìn)行了分析，給出了MPSO算法的粒子編碼、操作設(shè)計(jì)和混合優(yōu)化算法步驟。最后針對(duì)某超臨界600MW直流鍋爐高溫過(guò)熱器的過(guò)熱汽溫控制進(jìn)行了仿真試驗(yàn)，結(jié)果表明該方法具有更好的性能指標(biāo)。關(guān)鍵字。改進(jìn)PSO算法RBF神經(jīng)網(wǎng)絡(luò)優(yōu)化策略神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)控制過(guò)熱汽溫Neuralnetworkspredictivecont
3、rolforsuperheatedsteamtemperaturebasedonmodifiedparticleswarmoptimizationAbstractThesuperheatedsteamtemperatureofboilersuperheatedsysteminfiredpowerstationisatypicalobject，whichhasnonlinear．uncertain，largedelayandinertialcharacteristics．Atpresent．generalPIDcascadecontrol
4、systemismainlyadopted，butitcan’tworkwellwhenmoreinterferencecomeorcyclesarechanged．NeuraInetworkpredictivecontrolfNNPC)willtakefulladvantageofthenonlinear,self-organizing，self-learningperformanceofneuralnetworksandtherollingoptimizing，feedbackadjustingeffectivenessofpred
5、ictivecontr01．Itwillbemoresuitableforthisuncertainboilersuperheatedsystemcontr01．InordertoimprovetheperformanceofNNPC，bycombiningmodifiedparticleswarmoptimization(MPSO)wimNNPC．a(chǎn)modelpredictivecontrollerbased-onMPSOandradialbasisfunction(RBF)hybridoptimizationstrate：gY(MP
6、SO-RBF)ispresented。alsoanon．1inParoptimizationcontrollerbased-onM口Soispresented．Inallusiontotnesuperheatedsteamtemperaturecontrol，acascadecontrolsystembased-onneuralnetworkpredictivecontrolisconstructed．a(chǎn)lsoitspredictivemodel＆rollingoptimizingalgorithm＆feedbackadjusting＆
7、simulationparameterssettingproblemsareanalyzed，atthesaHi．etimepatticleencodedformatof～口SO&operatingdesignmethodandhybridoptimizationalgorithmsteparepresented．Finallythesimulationexpa-imentsaredoneforsuperheatedsteamtemperaturesystemcontrolofsuDel"critical600MWdircetcurre
8、ntbeiler,alsothismethodwillbeacom期sttoconventionaINNPC．thesimulationresultshaveshownthismethod’svalidit
