Abstract：For the problem of photovoltaic DC arc fault localization, this study investigates the electromagnetic radiation characteristics of the fault arc and proposes a location method based on a redundant antenna array and a weighted centroid algorithm. Firstly, the root mean square value of the electromagnetic signals collected by antennas during arc combustion is calculated, together with irradiance, are put into a BP neural network to predict the distance between the antennas and the arc. Then, a redundant antenna array is constructed to explore different quantities and layouts of antennas, selecting the antennas with the strongest received signals. The antenna coordinates and distance are input into the weighted centroid algorithm to obtain the positioning result. Finally, the K-means clustering algorithm is employed to improve the positioning accuracy. Experimental results demonstrate that the proposed method exhibits excellent localization capabilities.
林亮世, 高伟, 杨耿杰. 基于冗余天线阵列和加权质心算法的光伏系统直流电弧故障定位方法[J]. 电气技术, 2024, 25(4): 16-23.
LIN Liangshi, GAO Wei, YANG Gengjie. A direct current arc fault location method for photovoltaic systems based on redundant antenna array and weighted centroid algorithm. Electrical Engineering, 2024, 25(4): 16-23.
2024, Vol. 25 
