一种直流配电网电能质量扰动识别方法

打开文本图片集

引用格式：，，.一种直流配电网电能质量扰动识别方法[J].现代电子技术，2025，48（10）：118-126

关键词：直流配电网；电能质量；扰动识别；DBSCAN聚类；功率谱密度；核主成分分析；麻雀搜索算法；支持向量机中图分类号：TN86-34；TM711 文献标识码：A 文章编号：1004-373X（2025）10-0118-09

Abstract：With theincreaseof powerelectronicdevicesconnectedtothepowergrid，direct-current （DC）distribution networksaresuperiortotraditionaldistributionnetworksintermsofstrongtransmissonperformance，reducedlinelosses，and newenergyconsumption，graduallybecominganew trendin futuredistributiondevelopment.Toensurethestableoperationof the DC distributionnetworkandensure powerqualityanimproved sparowsearch algorithm-support vector machine （ISSA-SVM） powerqualitydisturbanceidentificationmethodbasedonkernelprincipalcomponentanalysis（KPCA）featuredimensionality reduction is proposed.Theformationmechanismsofvariouspowerqualityisueswerethoroughlyexplored，andsixfeatureswere extractedbasedonwaveformanalysis.TheDBSCANclusteringmethodisusedtodetectthepresenceofoutliersanddetermine whethertouseKPCAtoreducethedimensionalityoffeatures，enabling themtoachievegoodclusteringindiferentdata situations;The ISSAisusedtooptimizetheparametersofSVM，andtheSVMmodelisretrainedusingtheoptimizationresults. Theexperimentalresultsshowthattheproposedmethod has highaccuracyandcanefectivelyidentifypowerqualitydisturbance signals.

Keywords：DC distributiongrid;powerquality；disturance identification;DBSCANclustering;powerspectral density kernel principal component analysis;sparrow search algorithm;support vector machine

0 引言

随着新型电力系统的建设，相比于交流配电网而言，直流配电网具有灵活易控、传输损耗小，可提升电能质量、可良好地接受各种新型直流能源以及减少成本等优点[。（剩余9432字）