基于RNN⁃LSTM神经网络的小电流接地故障选线方法  

Line Selection Method of Low Current Grounding Fault Based on RNN⁃LSTM Neural Network

作　　者：朱晓红 杨伟荣 张蓉 张安斌 李盼盼 ZHU Xiaohong;YANG Weirong;ZHANG Rong;ZHANG Anbin;LI Panpan(Qujing Power Supply Bureau of Yunnan Power Grid Co.,Ltd.,Yunnan Qujing 655000,China;Hefei Yixin Electrical Science and Technology Co.,Ltd.,Hefei 230000,China)

机构地区：[1]云南电网有限责任公司曲靖供电局,云南曲靖655000 [2]合肥溢鑫电力科技有限公司,合肥230000

出　　处：《高压电器》2023年第7期213-220,共8页High Voltage Apparatus

基　　金：云南电网公司科技项目计划资助(YNKJXM20220143)。

摘　　要：针对配电网中小电流接地系统故障选线难的问题,提出一种基于RNN⁃LSTM神经网络的小电流接地故障多判据融合的选线方法。将各线路零序电流的基波特征分量、五次谐波特征分量、小波能量特征分量作为输入量,将故障线路作为输出量,建立小电流接地故障选线RNN⁃LSTM选线模型。最后用MATLAB/Simulink进行仿真验证,验证了算法的准确性和可行性。In view of the difficulty in selecting line in low current grounding systems in distribution networks,a kind of line selection method based on RNN⁃LSTM neural network is proposed for multi⁃criterion fusion of low current ground faults.The RNN⁃LSTM line selection model of the low current ground fault selection line is set up by taking the fundamental characteristic component,the fifth harmonic characteristic component and the wavelet energy char⁃acteristic component of the zero⁃sequence current of each line as the input and the fault line as the output.Finally,Matlab/Simulink is used for simulation verification,which verifies the accuracy and feasibility of the algorithm.

关 键 词：RNN⁃LSTM 小电流接地系统 故障选线 

分 类 号：TM862[电气工程—高电压与绝缘技术] TP183[自动化与计算机技术—控制理论与控制工程]