基于SSSC-EMD和BPNN的输电线路接地故障识别方法  

Ground Fault Identification Method for Transmission Line Based on SSSC-EMD and BPNN

作　　者：宋阳 武良 马驰 王玉帅 SONG Yang;WU Liang;MA Chi;WANG Yushuai(School of Electric Power,Shenyang Institute of Engineering,Shenyang 110136,Liaoning;Benxi Power Supply Company,State Grid Liaoning Electric Power Co.,Ltd.,Benxi 117000,Liaoning;Chongqing Research Institute,China Coal Technology Engineering Croup,Chongqing 400037;China Resources Power Tangshan co.,Ltd.,Tangshan 063000,Hebei)

机构地区：[1]沈阳工程学院电力学院,辽宁沈阳110136 [2]国网辽宁省电力有限公司本溪供电供电公司,辽宁本溪117000 [3]中煤科工集团重庆研究院有限公司,重庆400037 [4]华润电力唐山公司,河北唐山063000

出　　处：《济源职业技术学院学报》2024年第3期65-73,共9页Journal of Jiyuan Vocational and Technical College

基　　金：沈阳工程学院服务振兴技术创新研发先导项目(DLXDC2023004)。

摘　　要：本文报道了一种基于软筛分停止准则改进的经验模态分解(soft sifting stopping criterion-empirical mode decomposition,SSSC-EMD)和误差反馈传递型神经网络(back propagation neural network,BPNN)对输电线路接地故障进行识别的方法。通过在PSCAD/EMTDC中搭建输电线路模型,设置接地故障,将故障信号进行相模解耦后导入MATLAB中,分别使用SSSC-EMD+BPNN和EEMD+BPNN法对单相接地故障和两相接地故障进行识别对比。仿真结果表明:EEMD+BPNN法对单相接地故障的10组数据中有3组被识别为两相接地故障,对10组两相接地故障识别完全正确,故障识别的准确率为85%;而SSSC-EMD+BPNN法对单相接地故障和两相接地故障识别均正确,故障识别准确率达100%。This paper reports the soft sifting stopping criterion-empirical mode decomposition(SSSC-EMD)and error feedback neural network(BPNN)based on soft sifting stopping criterion-empirical mode decomposition(SSSC-EMD)and error feedback transmission type neural network(BPNN)for transmission line ground fault methods for identification,By building a transmission line model in PSCAD/EM’TDC,setting up ground faults,and importing the fault signals into MATLAB after phase-mode decoupling,single-phase ground faults and two-phase ground faults are identified and compared using the$SSC-EMD+BPNN and EEMD+BPNN methods respectively.Simulation results show that 3 out of 10 groups of data identified by the EEMD+BPNN method for single-phase ground faults are identified as two-phase ground faults,and the identification of 10 groups of two-phase ground faults is completely correct,and the accuracy rate of fault identification is 85%:And the SSSC-EMD+BPNN method identifies both single-phase ground faults and two-phase ground faults correctly,and the fault identification accuracy rate reaches 100%.

关 键 词：软筛分准则 模态分解 误差反馈机制 神经网络 接地故障 故障识别 

分 类 号：TM76[电气工程—电力系统及自动化]