Identification and Detection of DC Arc Fault in Photovoltaic Power Generation System

Abstract

This paper mainly studies the DC arc fault in photovoltaic system. First, the experimental platform of the arc fault of the photovoltaic system is set up, and the fault arc current signals under different conditions are collected. The time domain characteristics and the frequency domain characteristics are quantified to find out the time frequency characteristic of the arc. By analyzing the time domain characteristic of the fault arc, it is found that the system can be used as a criterion for identifying the fault arc in the time domain when the fault arc is occurring, the change of the working point, the peak of the current peak and the amplitude of the standard difference. Then the Flourier transformation is used to analyze the frequency components of the fault arc, and several frequency bands which have obvious changes in different conditions are used as the criterion of the frequency domain to identify the fault arc. In the end, 6 groups of eigenvectors are extracted as input of BP neural network, and the BP neural network is applied to the detection of DC fault arc. The validity and feasibility of this method are proved by experiments.