集成电路应用2024,Vol.41Issue(8) :277-279.DOI:10.19339/j.issn.1674-2583.2024.08.127

基于图卷积神经网络的配电网故障定位方法及算例分析

Analysis of Fault Location Method and Examples in Distribution Network Based on Graph Convolutional Neural Network

彭灵利
集成电路应用2024,Vol.41Issue(8) :277-279.DOI:10.19339/j.issn.1674-2583.2024.08.127

基于图卷积神经网络的配电网故障定位方法及算例分析

Analysis of Fault Location Method and Examples in Distribution Network Based on Graph Convolutional Neural Network

彭灵利1
扫码查看

作者信息

  • 1. 广东电网有限责任公司,广东 510180
  • 折叠

摘要

阐述为建立配电网故障的自动化定位方法,构建故障定位的算法模型.借助该算法从配电网的拓扑网络结构中提取节点特征和拓扑特征,将其作为模型输入,再经过数据标准化处理、图卷积层运算、全连接层运算,实现故障区段定位.检验结果显示,经过80次迭代后,此算法模型的定位精确度可以达到96%.

Abstract

This paper expounds the establishment of an automated fault location method for distribution networks and the construction of an algorithm model for fault location.It uses this algorithm to extract node and topology features from the topology network structure of the distribution network,takes them as model inputs,and then processes them through data standardization,graph convolutional layer operations,and fully connected layer operations to achieve fault section localization.The test results show that after 80 iterations,the positioning accuracy of this algorithm model can reach 96%.

关键词

图卷积神经网络/配电网/故障定位方法/算例分析

Key words

graph convolutional neural network/distribution network/fault location method/case analysis

引用本文复制引用

出版年

2024
集成电路应用
上海贝岭股份有限公司

集成电路应用

影响因子:0.132
ISSN:1674-2583
段落导航相关论文