The complex wavelet transform modulus maximum of the PD signal increases with scale, while the complex wavelet transform modulus maximum of white noise decreases with scale. According to the characteristics, a study on white noise suppression using the effective complex wavelet coefficient (ECWC) threshold method is launched in this paper and a comparison is conducted with the wavelet threshold denoising method of threshold selection of Stein unbiased risk estimate theory and threshold selection of minimax theory. The PD signal denoising results show that ECWC threshold method is more effective and the distortion of the extract PD signal is lower compared with the other method.

A numerical simulation for DC PD in void is put forward based on the PD physical process. The finite difference method is used to calculate the electric field distribution, and both of the stochastic property and the accumulation of the charge after PD on the void surfaces are considering in the model. The time of PD occurring, the amount of discharge and the voltage across the void are calculated. Meanwhile, the relationship between the DC voltage and the PD time interval or repetition rate is also simulation, the results show that with the increase of the DC voltage, the PD interval corresponding decreases exponentially and the repetition rate increases exponentially.