Partial Discharge Signal Denoising Analysis Using Wavelet Transformation with Singular Value Decomposition Method

Abstract

In high voltage power equipment, insulator role is important for reliability of the power system. Partial discharge (PD) analysis will help to determine the condition of insulator. However, in field due to interferences, it is difficult to detect PD signal without disturbances. PD signal with interferences may lead to misperceive of insulator condition. In this paper, a PD signal denoising method is proposed and its effectiveness are analyzed. The ability of discrete wavelet transformation (DWT) adopt to time localized and inhomogeneous signal has led to its recognition as an effective tool for PD processing. Discrete wavelet packet transform (DWPT) is variation of DWT, which gives more detailed analysis of signal. Singular value decomposition (SVD) is an algorithm that separates the singular values of the periodic narrowband interference (PNI) in PD signal and remove it. In this article, SVD with DWT and SVD with DWPT methods are proposed and applied to denoise the simulated and practical PD signals consisting of white noise interference (WNI) and PNI. The effectiveness of the methods used are determined by statistical analysis such as signal to noise ratio, root mean square, normalized correlation coefficient. From the results of the analysis, it is observed that, for denoising the PD containing WNI, DWPT method is effective, on the other hand to denoise the PD signal with both WNI and PNI, SVD with DWT is an effective method.