The Correlation Analysis of Photovoltaic Arc Fault Physical and Feature Parameters Based on the Mutual Information Entropy Method

At present, the photovoltaic (PV) arc fault feature model is mainly carried out from the circuit perspective. However, it cannot consider the influence of arc physical factors to verify the reliability of the arc fault detection algorithm. In this paper, the multi-dimensional arc fault feature model is first established to simulate PV feature parameters in the time domain and the frequency domain. Then, the MHD model of a dc air arc is established to obtain physical parameters including arc pressure, temperature, potential. Next, the physical and feature parameters of arc faults are obtained under the influence of different system current level and arc gap factors. Finally, the correlation model is established between physical and feature parameters based on the correlation analysis method. The correlation order of physical parameters is also acquired for multi-dimensional features respectively. These results explain the physical process of the influence of the system current and arc gap length on arc fault features in experimental PV systems. Furthermore, the arc feature simulation system combined with MHD also lay a foundation for the research of improving the reliability of the arc fault detection technology.