Single-phase Grounding Fault Location Method for Distribution Network Based on Convolutional Neural Network

Abstract

When a single-phase ground fault occurs in the distribution network, it is generally allowed to operate with faults from 1 to 2 hours, which may lead to further development of the fault and even threaten the safe operation of the power system. Therefore, when a small current system has a ground fault, it must be quickly diagnosed to shorten the time for the distribution network operation and maintenance personnel to eliminate the fault, thereby improving the safety of the distribution network operation. It proposes a single-phase grounding fault line selection method based on convolutional neural network (CNN). Phasor Measurement Unit (PMU) is used to extract the negative sequence voltage and current phasors at both ends of the fault line, and the principle of double-terminal fault location is applied to high-precision fault location. Simulink simulation results show that the proposed method can realize line selection and accurate location of single-phase ground fault in distribution network, and is not affected by system frequency, fault location, transition resistance or other factors.