Suppression of UHF partial discharge signals buried in white-noise interference based on block thresholding spatial correlation combinative de-noising method

摘要

White noise is a major noise that influences the accuracy of partial discharge (PD) measurements. According to the characteristics of white noise and PD signals, a new method for suppressing white noise when monitoring PD signal in gas insulated switchgear (GIS) is presented; this method combines respective advantages of block thresholding and spatial correlation de-noising arithmetic. First this algorithm uses a thresholding de-noising method to preprocess the sub-band signal of wavelet decomposition of noise, then processes the data with the spatial correlation method and finally reconstructs the de-noised information and returns it to the original signal. This method combines the respective advantages of the two de-noising methods. The impacts of decomposition level, noise intensity and iteration number of spatial correlation de-noising algorithm are discussed by the analysis result of simulation and experiment in the study. The simulative experimental results indicate that this method can extract PD signals effectively. Compared with complex Daubechies wavelet (CDW) transform adopted to suppress white noise, the new method has a better de-noising effect and less energy loss and performs more accurately. Finally, an example of a field-detected PD signal shows that the method is effective and practical in PD online monitoring.