Partial Discharge Localization of Gas-Insulated Switchgear via Electromagnetic Time Reversal Enhanced by Windowed Phase Spectrum Cross-Correlation Criterion

Abstract

Partial discharge (PD) localization is one of the most crucial steps in the insulation assessment of gas-insulated switchgear (GIS) since it can help technicians identify defective components efficiently. Recently, a new PD localization method named electromagnetic time reversal (EMTR) has been developed for GIS. The technique yields more accurate localization results than the conventional time-of-arrival and angle-of-arrival methods since it uses the three-dimensional structure information of GIS. However, the classical EMTR method analyzing individual sensor data independently suffers from the problems of multiple spurious solutions and noise interferences. Accordingly, this paper proposes an improved EMTR method for PD localization of GIS based on dual-sensor detection and phase spectrum cross-correlation analysis. The proposed method offers two improvements as follows: on the one hand, the problem of multiple spurious solutions is solved via the phase spectrum cross-correlation analysis between the electric field distributions in GIS stimulated by the two back-injected signals from the two sensors, respectively; on the other hand, a windowed phase spectrum cross-correlation approach is proposed to automatically select the frequency sub-bands with high signal-to-noise ratios for PD localization calculation, thus minimizing the influence of noise interferences. Finally, the feasibility and effectiveness of the proposed method are validated via the simulation experiment on a 110-kV GIS with a comprehensive comparison with state-of-the-art alternatives.

Keywords

• Electromagnetic time reversal (EMTR)
• partial discharge (PD) localization
• gas-insulated switchgear (GIS)
• cross-correlation