基于深度神经网络融合欧氏距离的多环配电网拓扑辨识方法  

Topology identification method for multi-ring distribution networks based on deep neural networks and Euclidean distance

作　　者：李博通[1] 孙铭阳 张婧[1] 陈发辉 陈晓龙[1] 王永祺 武娇雯[3] 魏然 LI Botong;SUN Mingyang;ZHANG Jing;CHEN Fahui;CHEN Xiaolong;WANG Yongqi;WU Jiaowen;WEI Ran(Key Laboratory of Smart Grid of Ministry of Education,Tianjin University,Tianjin 300072,China;North China Branch of State Grid Corporation of China,Beijing 100053,China;State Grid Tianjin Electric Power Company,Tianjin 300010,China)

机构地区：[1]天津大学智能电网教育部重点实验室,天津300072 [2]国家电网有限公司华北分部,北京100053 [3]国网天津市电力公司,天津300010

出　　处：《电力系统保护与控制》2025年第5期123-134,共12页Power System Protection and Control

基　　金：国家电网有限公司总部科技项目资助(5400-202412189A-1-1-ZN)“低压配电网环网供电模式与运行策略研究”。

摘　　要：针对多环配电网的拓扑辨识问题,考虑到量测信息可能部分缺失的情况,提出了基于深度神经网络融合欧氏距离的多环配电网拓扑辨识方法。首先,分析了传统拓扑辨识中相关性判断法应用于环状配电网的局限性,在此基础上提出基于欧氏距离的拓扑辨识判据。然后,针对量测信息缺失时的多环拓扑辨识问题,研究了利用深度神经网络融合欧氏距离判据的拓扑辨识方法。最后,在Matlab中利用MatPower搭建32节点“蜂巢”电网模型,在缺失不同比例的量测数据情况下验证方法的准确性。结果表明,当缺失大量量测数据时,所提方法仍有较高的拓扑辨识准确率。In response to the problem of topological identification for multi-ring power distribution networks and considering the possibility of partial loss of measurement information,a method for topological identification of multi-ring power distribution networks based on deep neural networks and Euclidean distance is proposed.First,the limitations of the traditional topological identification method using correlation judgment in ring-shaped power distribution networks are analyzed.Based on this,a topological identification criterion based on Euclidean distance is proposed.Then,to address the issue of topological identification of multi-ring networks with missing measurement information,a method combing deep neural networks with the Euclidean distance criteria is proposed.Finally,a 32-node“honeycomb”power grid model is built in Matlab using MatPower,and the accuracy of the method is verified under different levels of missing measurement data.The results show that even with a large amount of missing measurement data,the proposed method still maintains a high accuracy for topological identification.

关 键 词：欧氏距离 多环配电网 深度神经网络 拓扑辨识 量测信息缺失 

分 类 号：TM7[电气工程—电力系统及自动化]