Research on DC Arc Fault Detection Method Based on EWMA Dynamic Threshold

In recent years, arc faults caused by insulation aging and damage have become commonplace in photovoltaic power generation systems, and fault detection has become an indispensable part of photovoltaic power generation systems. DC arc faults have high-frequency characteristics, which are highly distinguishable for arc detection. Fast Fourier transform is used to extract their frequency domain characteristics, which are then compared with thresholds to determine whether a fault has occurred. The threshold setting is a key factor affecting the accuracy of detection. By collecting current data from inverters under different load conditions and arc fault locations, analyzing frequency domain characteristics, and using the frequency domain as a fault feature, we propose a method based on EWMA (exponential weighted moving average) frequency domain dynamic threshold detection. By validating the algorithm using arc data from different work scenarios, the dynamic threshold method can effectively detect arc faults in various scenarios.