基于图注意力网络的配电网故障定位方法

Distribution Network Fault Location Based on Graph Attention Network

基于人工智能的电网故障诊断技术已经有了大量的研究成果,但配电网拓扑变化频繁,而传统人工智能方法高度依赖训练数据,给配电网的故障定位问题带来了困难。提出了一种基于图注意力网络(graph attention network,GAT)的配电网故障定位方法。将配电网的电气节点和线路映射为图注意力网络中图的顶点和边,根据相邻顶点之间故障特征的相似度计算注意力系数,把顶点特征之间的相关性更好地融入到故障定位模型中,提高了故障定位模型对拓扑变化的适应能力。最后,搭建了配电网故障仿真模型验证了所提方法具有定位精度高、鲁棒性好且不受故障电阻、故障初相角和故障距离影响的优点,并在不同网络拓扑变化程度和情景下验证了模型在实际综合故障情景中有良好的应用效果。

There have been a lot of research achievements on the power system fault diagnosis technology based on artificial intelligence,but the topology of the distribution network changes frequently,and the traditional artificial intelligence method highly relies on the training data,which brings difficulties to the fault location of the distribution network.This paper proposes a fault location method for the distribution network based on the graph attention network(GAT).The electrical nodes and lines of the distribution network are taken as the vertices and edges of the graph in the graph attention network.The attention coefficient is calculated according to the similarity of the fault features between the adjacent vertices,and the correlation between the vertex features is better integrated into the fault location model,which improves the adaptability of the fault location model to the topology changes.Finally,this paper builds a distribution network fault simulation model to verify that the proposed method has the advantages of high positioning accuracy and good robustness,and it is not affected by fault resistance,fault initial phase angle and fault distance.Under different network topology changes and scenarios,the model has a good application effect in actual comprehensive fault scenarios.