第47卷第4期电网技术Vol.47No.42023年4月PowerSystemTechnologyApr.2023文章编号：1000-3673（2023）04-1612-11中图分类号：TM721文献标志码：A学科代码：470·40基于萤火虫扰动麻雀搜索算法极限学习机的光伏阵列故障诊断方法研究赵靖英1，吴晶晶1，张雪辉1，张文煜2，姚帅亮2（1．省部共建电工装备可靠性与智能化国家重点实验室(河北工业大学)，天津市北辰区300401；2．国网冀北张家口风光储输新能源有限公司，河北省张家口市075000）FaultDiagnosisofPhotovoltaicArraysBasedonSparrowSearchAlgorithmWithFireflyPerturbation-extremeLearningMachineZHAOJingying1,WUJingjing1,ZHANGXuehui1,ZHANGWenyu2,YAOShuailiang2(1.StateKeyLaboratoryofReliabilityandIntelligenceofElectricalEquipment(HebeiUniversityofTechnology),BeichenDistrict,Tianjin300401,China;2.StateGridHebeiZhangjiakouWindPowerStorageandTransmissionNewEnergyCo.,Ltd.,Zhangjiakou075000,HebeiProvince,China)ABSTRACT1:Withthecharacteristicsofrandomandintermittentoutputs,thefaultsoftheseriousphotovoltaicarraysaffectsthesafeoperationofthesystem.Inviewofthelackoffaultcharacteristicparametersthateffectivelycharacterizethedifferentdegreesofapartialshadingoragroundfaultintherainydays,theoperatingcharacteristicsofthephotovoltaicarraysunderdifferentfaultsisanalyzed,andanew6-dimensionalfaultfeaturevectorisproposed:TheopencircuitvoltageUoc,themaximumpowerpointvoltageUmandshortcircuitcurrentIsc,themaximumpowerpointcurrentImareintroducedtocharacterizetheshortcircuitandtheopencircuitfaultsrespectively.ThenumberofzeropointsofthesecondderivativeoftheU-Icurveisintroducedtocharacterizethepartialshading.Andthefuzzyc-meansclusteringalgorithmoptimizedbythegeneticsimulatedannealingalgorithm(GSA-FCM)isusedtoverifythevalidityofthedifferentdegreesofthepartialshadingcharacterizedbytheUmandIm.Thetotalharmonicdistortionrateofthegrid-connectedcurrentisintroducedtocharacterizethegroundfaultinrainydays.Thesparrowsearchalgorithmwithfireflyperturbation(FSSA)isintroducedtooptimizethetraditionalextremelearningmachine(ELM).AnFSSA-ELMmodelisestablishedtosolvetheproblemssuchascompleximplemen...
