Noise reduction method based on WOA-VMD-FE for fault arc acoustic signals in distribution networks

Arc acoustic signals are an effective means of characterizing faults in distribution networks, but their acquisition is often hindered by environmental limitations and various types of noise interference. To address the issue of misjudgment in fault detection due to noise and the challenge of selecting parameters for the Variational Mode Decomposition (VMD) algorithm, this paper proposes a denoising method based on the Whale Optimization Algorithm (WOA) for arcing acoustic signals in distribution network faults. Firstly, the Whale Optimization Algorithm is used to optimize the penalty factor α and the decomposition modulus K of VMD, thereby determining the optimal decomposition parameters. Secondly, fuzzy entropy is employed to evaluate each Intrinsic Mode Function (IMF) signal decomposed by VMD, enabling the distinction between the noisy components and the effective signal parts. The Sparrow Search Algorithm, Particle Swarm Optimization, Genetic Algorithm, and the Whale Optimization-based VMD parameter searching method are compared. The results demonstrate that the proposed method significantly improves the accuracy and speed of VMD parameter optimization compared to the Sparrow, Particle Swarm, and Genetic Algorithms. Effectively filtering out ambient noise while preserving the original characteristics of the signal to a great extent.