Noise removal from power system signals

Abstract

This paper proposes a method for noise removal from power system disturbance signals based on wavelet packet transform (WPT). One of the major problems in detecting disturbances in power system signals is the presence of noise. To enhance the standard WPT technique in processing the noise-riding signals, this paper proposes a noise removal algorithm. The denoising of the signals is performed by applying the proposed weighted Shannon entropy and the level dependent minimum description length (MDL) algorithms. Weighted Shannon entropy is employed for the best basis selection from the complete wavelet packet tree and the level dependent MDL is utilized for the thresholding of the coefficients at each node of the optimal sub-tree. The proposed method is evaluated using a set of simulated and real power system signals with disturbances in Smart Grid. The results indicate that level dependent MDL based WPT method not only depress the noise contained in the signals but also compress the disturbance signals. The performance of the proposed method is compared with Shannon entropy and the standard MDL.