GIS Partial Discharge Location Method Based on UHF Fingerprint and CAPSO-FNN

Achieving efficient detection and localization of partial discharge in gas insulated switchgear is very important for timely detection of insulation defects and improvement of maintenance efficiency. The identification and positioning of partial discharge are crucial factors in monitoring and diagnosing the condition of power equipment in substations. Due to its expensive equipment cost, the current application scope of the time difference-based ultra-high frequency partial discharge location technology is restricted. Addressing the challenge posed by the extended axial distance of the GIS air chamber and the complexity of locating the partial discharge source, a location method based on ultra-high frequency signal simulation fingerprint and CAPSO-FNN is proposed, which can realize the accurate location of partial discharge source. This method introduces ultra-high frequency simulation data into PD source location, which overcomes the problem that conventional fingerprint-based location methods need to collect a large number of field experimental data. By establishing a simulation model with the same size as the experimental GIS, the ultra-high frequency simulation signals of partial discharge sources at different locations are obtained, and the positioning simulation fingerprint database containing coordinate information is constructed. Subsequently, the CAPSO algorithm was used to optimize the FNN model, and the optimized CAPSO-FNN model was used to pattern match the measured partial discharge fingerprint with the fingerprint database, obtaining the corresponding spatial coordinates of the partial discharge source. According to the experimental findings, the average positioning error of this method is below 0.48cm, which can achieve precise positioning of partial discharge sources in GIS, and the positioning effect is significantly better than conventional BP algorithm and ELM algorithm.