基于CNN的配电网接地故障识别  

Grounding Fault Identification of Distribution Network Based on CNN

作　　者：李宇 LI Yu(State Grid Sichuan Electric Power Company Guang′an Power Supply Company,Guang′an 638500,China)

机构地区：[1]国网四川省电力公司广安供电公司,四川广安638500

出　　处：《电工技术》2023年第17期83-85,89,共4页Electric Engineering

摘　　要：对配电网不同类型接地故障的准确识别有助于提高配电网供电可靠性。鉴于配电网发生接地故障时,故障信息微弱且分类器挖掘故障特征能力有限,提出了时分频分方法提取故障特征,基于卷积神经网络(CNN)实现不同接地故障分类。对仿真实验结果的分析表明,相较于传统机器学习方法SVM、KNN、DT,所提方法的评估指标结果更优。Accurate identification of different types of grounding faults in distribution network is helpful to improve the reliability of power supply.Grounding faults in distribution network has weak fault information,and the ability of current classifiers for mining fault characteristics is insufficient.In view of the aforementioned problem,a time-division and frequency-division method is proposed to extract fault characteristics,and the classification of different grounding fault is realized based on convolution neural network(CNN).Simulative analyses show that the evaluation index results of the proposed method are better than those of traditional machine learning methods such as SVM,KNN and DT.

关 键 词：配电网 故障识别 卷积神经网络 小波分析 

分 类 号：TM726[电气工程—电力系统及自动化]