Search

Partial Discharge (PD) will make insulation of electrical apparatus decrease and threaten the security and stability of power system. On the other hand, it is also an indicator of insulation status of electrical apparatus. Therefore, it is important to monitor PD online. However, inevitable white noise exists and will interfere PD measurement, thus distorting it and leading to misjudging of insulation status. Therefore, denoising of PD signal is very necessary. Wavelet thresholding denoising method is one of the most commonly used method, but it lacks translation in-variance, and the denoising effect is influenced. To overcome this disadvantage, common factor method is proposed to construct dual-tree complex wavelet combined with wavelet thresholding in this paper. Finally, numeric simulation experiments have been set to verify the feasibility and effectiveness of the proposed method.

In high voltage engineering, partial discharge (PD) is an important topic. PD measurement is frequently used to monitor the status of critical equipment such as power transformer, gas insulated switchgear, and cable. The measured PD signal is of great importance for insulation evaluation and fault diagnosis. However, PD signal easily gets corrupted by periodic narrow-band noise, which will cause interference and finally lead to misjudgment of insulation status. Therefore, it is necessary to denoise it in the measured PD signal. In this paper, time-frequency analysis is proposed for denoising. Instantaneous frequency (IF) is also introduced. Due to the different IF characteristics between pure PD signal and periodic narrow-band noise, it is possible to distinguish and separate them. Numeric simulation experiments have been carried out to verify the feasibility of the proposed method.