A Series DC Arc Fault Detection Algorithm Based on PV Operating Characteristics and Detailed Extraction of Pink Noise Behavior

Traditional arc fault detection methods focus on the time domain or the frequency domain of the current signal as an indicator of a possible arc fault event. However, the pink noise generated by the arc fault is usually observed only as an increase in energy and not specifically characterized, leading to unwanted tripping from other arc-like transients. In this research, the pink noise analysis, along with unique properties of the PV operating point of both the instantaneous current and voltage signal, is used to distinguish arc faults from other noisy loads or PV degradations. Experimental results validate that the algorithm successfully detects series DC arc faults.