Ground Fault Localization of Branched Wire Network using Reverse Image Search

For the branched wire network, the identification and location of ground fault are difficult when implementing the reflectometry method since multiple reflections can be from different branches and reflected signal peaks can overlap with each other. In this paper, a novel branched network ground fault location method using the time-frequency domain reflection (TFDR) method and reverse image search is proposed and tested through simulation and experiment. The proposed method is tested using Y-shape branched network. A ground fault scenario database is built using simulation and the experimental ground fault Wigner-Ville distribution figure is compared with the database to find the most similar figure. Experimental results on two fault scenarios demonstrate the effectiveness of the proposed methods.