Detection of Series Arc Faults in Photovoltaic System Based on Wavelet Filter Evaluation

Series Arc Faults (SAFs) in PV systems are very difficult to detect owing to inverter noise. Therefore, this study aims to detect SAFs by evaluating data collected from three different measurement: current measurement, voltage measurement on the arc fault generator, and PV voltage measurement with three different case studies based on the amount of Photovoltaic (PV) voltage from the Direct Current Power Supply (DCPS) that can simulate real PV conditions. Three wavelet filter types, such as Haar, Daubechies 4 (db4), and Biortogonal 3.1 (3.1) with a decomposition level of 15. The findings of this research indicate that the PV voltage is capable of measuring SAFs with evaluating wavelet filter. In addition, the arcing power in the steady state increased as the PV source voltage increased. The threshold limits for normal conditions can be established using the Haar wavelet filter method and db4 detail coefficient value. However, it is not recommended to utilize the Bior 3.1 method for detecting SAFs in PV systems because of the high inverter noise during normal conditions in some cases. The contribution of this study is that detection using the PV voltage can be performed for arcing detection with wavelet filter evaluation.