A Fault Detection and Location Method by Applying the Attenuation Law of Arc Signal Conduction

In recent years, the detection of arc fault in photovoltaic (PV) system has become a hot issue. Dc arc faults may pose serious threats to personal safety and property. In this paper, a fault location method considering the attenuation rule of signal conduction under the conduction coupling interference of PV system is studied. Based on wavelet transform (WT), time-frequency conversion of arc fault current data is carried out. In combination with random forest (RF) model, arc fault state identification of PV system under conductive coupling interference is realized. Finally, the random forest regression (RFR) algorithm is used to locate the arc fault.