基于图神经网络的分布式电源接入配网故障区段在线定位方法

Online Positioning Method of Fault Section of Distributed Power Supply Access Distribution Network Based on Graph Neural Network

由于传统故障区段定位方法定位结果误差较大,提出基于图神经网络的分布式电源接入配网故障区段在线定位方法。利用电源管理单元(Power Management Unit,PMU)同步测量分布式电源接入配网节点的电压信号、电流信号、零序分量等电气量数据,经过一系列预处理,构建一个图神经网络模型,学习节点电气量数据与故障区段之间关系,实现分布式电源接入配网故障区段在线定位。实验结果表明,应用所提方法,可精准定位故障区段,且耗时较短,应用效果较好。

Due to the large error in the positioning results of traditional fault section positioning methods,a distributed power supply access distribution network fault section online positioning method based on graph neural network is proposed.The article uses Power Management Unit(PMU)to synchronously measure electrical quantity data such as voltage signals,current signals,and zero sequence components of distributed power sources connected to distribution network nodes.After a series of preprocessing,a graph neural network model is constructed to learn the relationship between node electrical quantity data and fault sections,achieving online positioning of fault sections in distributed power sources connected to distribution networks.The experimental results show that the proposed method can accurately locate the fault section,with short time consumption and good application effect.