摘要
目前运用配电网中的多源故障遥信数据对配电网故障进行诊断的技术逐渐火热,但由于信息的误报、漏报使得诊断结果往往不尽如人意。针对配电网多源故障信息不健全的问题,本文提出了一种将数据校核操作与图卷积神经网络相结合的高容错性配电网故障诊断方法,旨在利用不健全故障信息对配电网进行故障诊断。首先,将多源故障遥信数据作归一化处理后再进行数据校核,从源头改善数据的不健全性;接着,根据配电网图模将多源故障数据转化为故障图数据;最后,将故障图数据送入图卷积神经网络进行学习训练,训练完毕的模型在部署后可实现对配电网的故障诊断。在Python 3.7平台进行实验,通过算例分析证明本文所提方法可有效提高配电网故障诊断的容错性。
At present,the technology of using multi-source fault remote signaling data to diagnose the fault of distribution network is becoming increasingly popular,but the diagnosis results are often unsatisfactory due to the false alarm and missing report of data.Aiming at the problem of imperfect multi-source fault information,this paper presents a fault diagnosis method for distribution network with high fault tolerance,which combines the data check operation with graph convolution neural network.It aims to use the imperfect fault information to diagnose the fault in distribution network.Firstly,and the data after normalization of multi-source fault data are checked,so as to improve the soundness of the data from the source;then,transform multi-source fault data into fault diagram data are transformed according to distribution network diagram model;and finally,the fault diagram data are sent to the graph convolutional neural network for learning and training,and the trained model can realize fault diagnosis for distribution network with high fault tolerance after deployment.The simulation experiment on Python 3.7 platform shows that the proposed method can effectively improve the fault tolerance of distribution network fault diagnosis.
作者
高艺文
苏学能
张华
姜思远
高红均
GAO Yiwen;SU Xueneng;ZHANG Hua;JIANG Siyuan;GAO Hongjun(Electric Power Research Institute,State Grid Sichuan Electric Power Company,Chengdu 610041,China;College of Electrical Engineering,Sichuan University,Chengdu 610065,China)
出处
《电工电能新技术》
CSCD
北大核心
2024年第2期95-104,共10页
Advanced Technology of Electrical Engineering and Energy
基金
国网四川省电力公司科技项目(52199722000Q)。
关键词
多源故障数据
信息不健全
配电网故障诊断
数据校核
图卷积神经网络
高容错性
multiple fault data
imperfect information
distribution network fault diagnosis
data check
graph convolution neural network
high fault tolerance
