Noise Reduction Study of Pressure Pulsation in Pumped Storage Units Based on Sparrow Optimization VMD Combined with SVD

Abstract

The unbalanced forces generated by pumped storage units operating under non-ideal operating conditions can cause pressure pulsations. Due to the noise interference, the feature information reflecting the operating state of the unit in the pressure pulsation is difficult to extract. Therefore, this paper proposes a noise reduction method based on sparrow search algorithm (SSA) optimized variational mode decomposition (VMD) combined with singular value decomposition (SVD). Firstly, SSA is used to realize the adaptive optimization of VMD parameters for ideal decomposition of the signal. Then, the noise reduction of the decomposed signal is performed by using the sensitivity of the Permutation Entropy (PE) for small mutations. The noise reduction and reconstruction of the decomposed signal are carried out again by using SVD. The experimental and comparison results show that the mean square error of the signal after VMD-SVD feature extraction is reduced from 1.0068 to 0.0732 and the correlation coefficient is increased from 0.2428 to 0.9614. It is proved that the method achieves better results in the pressure pulsation signal of pumped storage units and has some application significance for the fault diagnosis of pumped storage units.