Proposal of a DC Arc Fault Location Identification Method for Application to Photovoltaic Systems

The widespread deployment of photovoltaic systems (PVS) is progressing rapidly. However, as the scale of these systems increases, two major challenges have emerged: the elevated fire risk due to arc faults and the difficulty in locating such faults across large installations. Although conventional arc fault circuit interrupters (AFCIs) are capable of preventing fires by disconnecting the circuit, they do not provide fault location information. This necessitates time-consuming manual inspections and results in extended system downtime. This study proposes a novel method to address this issue. Based on experimental investigations into the noise characteristics of arc faults, we identified a specific high-frequency band in which the signal amplitude exhibits an attenuation with respect to the propagation distance from the fault origin. Utilizing this characteristic, we developed a localization technique employing an analog circuit, which includes a band-pass filter to extract the target frequency component and an amplifier to process its amplitude. The fault location is estimated from the resulting output voltage. The proposed method enables rapid and accurate localization of arc faults, thereby reducing maintenance time and costs, and contributing to improved safety and operational reliability of large-scale PVS installations.