A Denoising Method for Partial Discharge Based on Wavelet Transformation and Singular Value Decomposition

Abstract

PD detection is a crucial technique used to assess the condition of power equipment's insulation. However, the PD signal measured in the field is typically prone to noise interference, which can significantly impact the accuracy of the detection results. To mitigate the effects of noise, a novel noise suppression method has been proposed, which utilizes both the discrete wavelet transform (DWT) and singular value decomposition (SVD). Firstly, the original signal containing four PDs is generated by simulation, and then superimposed with Gaussian white noise. Then, the whole PD signal mixed with white noise is decomposed and denoised by SVD, and the segments containing PD signals are accurately found by using different decomposition layers of DWT, and the segments of non-PD signals are zeroed. Then, SVD is performed on the segments containing only PD signals to obtain the PD signal after noise reduction. Finally, various methods have been employed to denoise partial discharge analog signals, and their efficacy has been assessed using different indicators to validate the effectiveness of our proposed approach.