Research on Time Delay Estimation Algorithm of GIS Partial Discharge Signal Based on Higher-Order Cumulant and Bispectrum Estimation

Partial discharge (PD) frequently occurs in high-voltage electrical apparatus, like Gas Insulated Switchgear (GIS), potentially causing the deterioration and damage of insulation materials, which can compromise the equipment's safety and stability. Consequently, effective monitoring and localization of PD are crucial. This study focuses on a time delay estimation method for GIS PD signals that leverages higher-order cumulants (HOC) and bispectrum analysis. Traditional techniques for time delay estimation tend to perform poorly under low signal-to-noise ratio (SNR) conditions. In contrast, HOC-based approaches excel at suppressing Gaussian noise and highlighting the non-Gaussian features of the signals, making them more suitable for such challenging environments. Bispectral estimation reveals the third-order statistical characteristics of the signal and provides rich information for signal analysis. Through theoretical analysis and experimental verification, this study shows the potential of these two methods to improve the accuracy and stability of time delay estimation under low SNR conditions. The algorithm framework includes signal preprocessing, ad HOC calculation, bispectrum analysis and time delay estimation. The simulation experiment was carried out in a professional laboratory, using simulated GIS cavity and actual GIS equipment signals. The results show that the proposed algorithm has significantly higher time delay estimation accuracy and better stability under low SNR conditions. This study provides a new method for time delay estimation of GIS PD signal, which is of great significance to improve the monitoring and positioning accuracy of GIS equipment.