A Fault Location Method Considering Distribution Network Partition Based on Deep Learning

Abstract

In this paper, a fault location method considering distribution network partition based on deep learning is proposed, in which the Tensorflow framework is employed to establish and construct the fault location model of the distribution network. This method firstly collects the current and voltage data to form fault data vectors through the Feeder Terminal Unit. Combined with the complex network theory, each node degree is calculated to represent the node priority, and the topology of the distribution network is partitioned to form each regional model. Secondly, it builds a feature extracting network and a Deep Neural network to mine the mapping relations between fault data vectors and fault sections and form the final fault location model through training. Case studies show that compared to the back propagation (BP) neural network model and the support vector machine (SVM) model, the deep learning model has faster convergence speed and higher fault location accuracy.