An innovative photovoltaic DC arc fault detection method through multiple criteria algorithm based on a new arc initiation method

The DC arc fault is a great threat to the safety of PV systems. Great heat from sustained arcs leads to fire accidents and casualties. As a result, the US 2011 National Electrical Code requires that PV systems shall be protected by a DC arc-fault circuit interrupter. In this paper, a new arc initiation method is used for experiment data collection instead of the conventional method mentioned in UL 1699B. The proposed method of arc ignition by separating electrodes controlled by stepping motor gains more stability and repeatability than that using fine steel wool and tube. Through analyzing large amounts of experiment data collected by this means, a DC arc detection method based on multiple criteria is proposed and corresponding thresholds are configured. This multiple criteria algorithm decreases the rate of false detection dramatically with both time and frequency domain arc characteristics considered. In the future, further work like developing prevention strategies and prototype testing will also be based on the very initiation method and detection algorithm.