A Novel Denoising Method for Ultra-High Frequency Signals of Partial Discharge in GIS/GIL

Partial discharge (PD) detection is an important means for insulation condition monitoring. The ultra-high frequency (UHF) method is commonly used for on-site PD detection in gas-insulated switchgear (GIS) and gas-insulated transmission line (GIL). However, the obvious attenuation of UHF signal and complex interference result in a low signal-to-noise ratio of PD signals, making it difficult to locate PD source. In this paper, a novel denoising method based on multi-channel correlation-average for UHF PD signals is proposed and displayed. Then, PD detection and localization in a hydropower plant is conducted. The results show that the UHF PD signal (about 5 mV) submerged in noise (about 20 mV) can be effectively extracted using the proposed denoising method. Subsequently, based on the denoised signal, the PD source was accurately located. This further validates the effectiveness of the proposed method. The research of this study provides good guidance for on-site PD detection and equipment maintenance.