Denoising of Cable Partial Discharge Signals Based on IACEEMDN and Improved Wavelet Thresholding

In response to the problems of periodic narrow-band interference, white noise, and the measurement challenges of cable partial discharge (PD) signals, this paper proposes a denoising method based on Improved Adaptive Complete Ensemble Empirical Mode Decomposition with Noise (IACEEMDN) and advanced wavelet threshold denoising. Specifically, the IACEEMDN algorithm decomposes the signal, and we identify effective Intrinsic Mode Functions (IMFs) using a method based on correlation coefficients. These effective IMFs are then denoised with an enhanced wavelet threshold algorithm. Subsequently, the denoised IMF components are reconstructed to yield the final denoised signal. Simulations and experimental validations indicate that the method put forth in this study more effectively removes noise and preserves the original characteristics and detailed parameters of the signal compared to conventional approaches.