Application of CEEMD and permutation entropy in noise elimination of hydropower unit swing signal

Abstract

A noise elimination algorithm of hydropower unit swing signals by using complete ensemble EMD (CEEMD) and permutation entropy is presented. The CEEMD is selected to decompose hydropower unit swing signals into several intrinsic mode functions (IMFs). Each IMF’s permutation entropy is computed. These IMFs (whose permutation entropy is lower than preset threshold) are reconstructed to denoise these abnormal noises of hydropower unit swing signals. Wavelet transform and CEEMD are used to process noise of hydropower unit swing signals respectively. Noise reduction performance of two methods is compared. The results can be seen that the CEEMD method has better adaptability and denoising capacity. It’s perfect for online applications for de-noising hydropower unit swing signals.