基于KNN算法的配电网单相接地故障选线研究

Research on Single Phase to Ground Fault Line Selection of Distribution Network Based on KNN Algorithm

小电流接地系统单相接地故障选线是配电网领域的一个难题,针对传统的采用单一判据的故障选线方案适用性差、选线精度低的问题,提出了一种基于K近邻(K-nearest-neighbor,KNN)算法的多源信息融合的单相接地故障选线方法,通过对故障数据处理选取故障特征量,结合KNN算法进行故障线路选线.算例仿真研究表明,该选线方法与逻辑回归算法、BP神经网络算法相比,在获得较高的准确率的同时可缩短选线时间,具有较好的应用前景.

Single phase to ground fault line selection is a difficult problem in the field of distribution network.In view of the poor applicability and low line selection accuracy of traditional fault line selection schemes that use a single criterion,this paper proposes a single phase to ground fault line selection method based on K-nearest-neighbor(KNN)algorithm combined with multi-source information fusion,which selects the fault feature quantity for the fault data processing and combines the KNN algorithm to select the fault line.The simulation study of the calculation example shows that the line selection method can obtain higher accuracy and significantly reduce the line selection time compared with the logistic regression algorithm(LoR)and the back propagation neural network algorithm(BP),and that it has a better application prospect.