Abstract: The Variational Mode Decomposition (VMD) algorithm has been used extensively for denoising swing signals from hydroelectric power generation units; it effectively addresses the issue of modal mixing encountered in the traditional algorithm through modal decompositions. However, its modal decomposition number and penalty factor must be optimized to achieve the best accuracy. This paper presents a fusion HO-VMD algorithm for denoising these unit swing signals to overcome the limitations of slow parameter selection and poor generalization capability found in previous VMD algorithms. This new method adopts the Hippopotamus Optimization (HO) algorithm to optimize both the number of modal decompositions and the penalty factor, and its performance is compared with that of the Improved Sparrow Optimization Algorithm (SSA) and Wavelet Transform VMD Algorithm. The results demonstrate this fusion HO-VMD algorithm is effective and suitable for denoising swing signals from hydroelectric power generation units.

Key words: hydroelectric generator, vibration signal, noise reduction, VMD, HO