Research On The Method Of Partial Discharge Interference Suppression

Partial discharge(PD)is often caused by harmful electrical discharge in the inner or surface of the insulator due to the excessive concentration of the electric field.Its existence seriously affects the insulation reliability of the electrical equipment.Therefore,partial discharge detection is an effective method to judge the insulation performance of electrical equipment.However,in the detection process,the collected partial discharge signals are usually subject to strong electromagnetic interference,such as white noise,narrowband noise,etc.So,the partial discharge signal interference suppression is an important link of PD detection.Partial discharge signals are usually collected by sensors,and then converted to digital signals via A/D.Therefore,the research on noise suppression of PD is mainly focused on the digital anti-jamming algorithm.This paper focuses on the digital filtering in PD signals.Wavelet transform is an important tool for signal analysis.It has wide applications due to its better localization in both time and frequency domains.Wavelet threshold de-noising is a mainstream signal processing method.Based on the analysis of the difference between the PD signal and the white noise in the wavelet domain,wavelet threshold de-noising is improved in this paper.First of all,in the choice of wavelet base,this paper uses the hierarchical optimal wavelet selection algorithm to further promote the separation of signal and noise;Secondly,for the problem of wavelet threshold,a new threshold function is proposed,and the quality of the threshold function is analyzed specially.The new threshold function can effectively improve the shortcomings of the conventional hard threshold and the soft threshold mode.The simulation results show that the improved wavelet threshold filtering method can effectively improve the extraction effect of PD signals.It is found that the wavelet threshold method has some limitations in the suppression of narrowband noise.Although FFT filtering has the ability to suppress narrowband interference,the choice of threshold often limits its application.In addition,the computational complexity has to be considered in the case of large numbers of sampling points.In order to eliminate the narrow band interference in PD signal effectively,a simple and practical digital trap is introduced in thispaper.The frequency domain characteristics of notch filter are analyzed emphatically.Finally,the simulation results show that the notch filter can suppress narrowband noise effectively.