| 引用本文: | 何小龙,高红均,黄 媛,等.基于一维卷积和图神经网络的配电网故障区段定位方法[J].电力系统保护与控制,2024,52(17):27-39.[点击复制] |
| HE Xiaolong,GAO Hongjun,HUANG Yuan,et al.Fault section location for a distribution network based on one-dimensional convolution and graph neural networks[J].Power System Protection and Control,2024,52(17):27-39[点击复制] |
|
| 摘要: |
| 快速、准确地定位故障区段对配电网的安全运行至关重要。传统故障定位方法容错率低、耗费时间长，多数深度学习算法对拓扑变动的泛化性不足。基于此，提出了一种基于一维卷积神经网络(one-dimensional convolutional neural network, 1D-CNN)和图神经网络(graph neural network, GNN)的配电网故障区段定位方法。该方法将配电网原始信息与GNN等深度学习算法相结合进行建模。首先利用基于注意力的时空图卷积网络从不同的时空尺度上对遥测数据进行故障特征提取，使用图注意力网络来融合多源遥信数据。然后，利用1D-CNN来调整特征输出维度以实现节点特征到故障支路的映射。最后，通过增设全连接网络来输出故障区段定位结果。依托于Matlab/Simulink平台搭建10 kV中性点不接地配电网系统进行仿真和测试。结果表明，所提方法具有优越的定位性能，能够灵活适用于各类低、中、高阻性接地故障场景，对系统拓扑变动具有强大的泛化能力以及对故障数据不完备的鲁棒性好。 |
| 关键词: 配电网 故障区段定位 一维卷积 图神经网络 拓扑变动 数据不完备 |
| DOI：10.19783/j.cnki.pspc.240021 |
| 投稿时间：2024-01-08修订日期：2024-05-16 |
| 基金项目:国家自然科学基金项目资助(52077146)；四川省科技计划项目资助(2023NSFSC1945) |
|
| Fault section location for a distribution network based on one-dimensional convolution and graph neural networks |
| HE Xiaolong1,GAO Hongjun1,HUANG Yuan1,GAO Yiwen2,WANG Renjun1,LIU Junyong1 |
| (1. College of Electrical Engineering, Sichuan University, Chengdu 610065, China;
2. State Grid Sichuan Electric Power Company, Chengdu 610041, China) |
| Abstract: |
| Quickly and accurately locating the fault section is crucial for the safe operation of a distribution network. Traditional fault location methods have low fault tolerance and take a long time, and most deep learning algorithms have insufficient generalization to topological changes. Thus a fault section location method for a distribution network based on a one-dimensional convolutional neural network (1D-CNN) and a graph neural network (GNN) is proposed. This method combines the original information of the distribution network with deep learning algorithms such as GNN for modeling. First, an attention-based spatial-temporal graph convolutional network (ASTCGN) is used to extract fault features from telemetry data at different temporal and spatial scales. The graph attention network (GAT) is used to fuse multi-source remote signaling data. Then, 1D-CNN is used to adjust the feature output dimension to realize the mapping of node features to fault branches. Finally, the fault section location results are output by adding a fully-connected network (FCN). A 10 kV neutral ungrounded distribution network system for simulation and testing is built based on the Matlab/Simulink platform. It is verified that the proposed method possesses superior positioning performance, can be flexibly applied to various low, medium, and high resistance grounding fault scenarios, and has strong generalizability to system topology changes and good robustness to incomplete fault data. |
| Key words: distribution network fault section location 1D-CNN GNN topology change incomplete data |
