資源描述:
《基于改進(jìn)foa優(yōu)化bp神經(jīng)網(wǎng)絡(luò)算法的》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、第44卷第504期電測(cè)與儀表Vol.44No.5042017年第12期ElectricalMeasurement&InstrumentationDec.2017基于改進(jìn)FOA優(yōu)化BP神經(jīng)網(wǎng)絡(luò)算法的光伏系統(tǒng)MPPT研究*閆超1,2,倪福佳1,劉嘉瑜1,2,賀詩(shī)明2,高振遠(yuǎn)2,王少帥1(1.中國(guó)礦業(yè)大學(xué)江蘇省煤礦電氣與自動(dòng)化工程實(shí)驗(yàn)室,江蘇徐州221116;2.中國(guó)礦業(yè)大學(xué)電氣與動(dòng)力工程學(xué)院,江蘇徐州221116)摘要:針對(duì)基于BP神經(jīng)網(wǎng)絡(luò)的光伏系統(tǒng)MPPT策略在光照強(qiáng)度突變時(shí)存在較大誤差的問(wèn)題,本文提出了一種改進(jìn)的果蠅優(yōu)化算法用于BP神經(jīng)網(wǎng)絡(luò)的權(quán)值和
2、閾值優(yōu)化,并建立了基于IFOA-BP神經(jīng)網(wǎng)絡(luò)算法的光伏系統(tǒng)MPPT控制的仿真模型。測(cè)試和仿真結(jié)果表明,IFOA的收斂速度和求解精度較改進(jìn)前均有明顯提升;IFOA優(yōu)化后的BP神經(jīng)網(wǎng)絡(luò)收斂速度加快,預(yù)測(cè)誤差減少;較之于電導(dǎo)增量法,IFOA-BP神經(jīng)網(wǎng)絡(luò)的MPPT策略在穩(wěn)態(tài)條件下能明顯抑制功率波動(dòng),在外界條件發(fā)生突變時(shí),能迅速準(zhǔn)確地追蹤到最大功率點(diǎn),具有良好的穩(wěn)態(tài)精度和動(dòng)態(tài)特性。關(guān)鍵詞:光伏電池;最大功率點(diǎn)跟蹤;BP神經(jīng)網(wǎng)絡(luò);改進(jìn)果蠅優(yōu)化算法中圖分類號(hào):TM933文獻(xiàn)標(biāo)識(shí)碼:B文章編號(hào):1001-1390(2018)00-0000-00Researcho
3、fonthephotovoltaicsystemMPPTbasedonimprovedIFOA-BPneuralnetworkalgorithmYanChao1,2,NiFujia1,LiuJiayu1,2,HeShiming1,GaoZhenyuan1,WangShaoshuai1(1.JiangsuProvinceLaboratoryofElectricalandAutomationEngineeringforCoalMining,ChinaUniversityofMining&Technology,Xuzhou221116,Jiangsu,C
4、hina.2.SchoolofElectricalandPowerEngineering,ChinaUniversityofMining&Technology,Xuzhou221116,Jiangsu,China)Abstract::WhentheBPneuralnetworkisadoptedtopredictthevoltageatthemaximumpowerpoint,thereisabigerrorifthelightintensitychangesdrastically.Aimingatthisproblem,anovelimprove
5、dfruitflyoptimizationalgorithm(IFOA)determiningtheoptimalBPneuralnetworkparameters(weightandthreshold)isproposed,andasimulationmodelofthephotovoltaicsystemMPPTcontrolstrategybasedontheIFOA-BPneuralnetworkalgorithmwasisestablished.Thetestandsimulationresultsshowthatthe,IFOAhasa
6、greatadvantageinconvergencesearchspeedandsolutionaccuracythanFOA;IFOA-BPneuralnetworkcaneffectivelyincreasestheconvergencespeedandreducesthepredictionerror;.comparedComparedwiththeincrementalconductance(INC)method,theproposedphotovoltaicsystemMPPTcontrolalgorithmbasedonIFOA-BP
7、neuralnetworkcouldsuppresstheoscillationaroundthemaximumpowerpoint(MPP)understeady-stateconditionsandtrackdowntheMPPquicklyandaccuratelywhenlightintensityandtemperaturechangedrastically,whichverifiesthestability,precisionandrapidityoftheproposedMPPTmethod.Keywords::photovoltai
8、ccell,maximumpowerpointtracking,BPneuralnetwork,improvedfruit