Denoising Method Of Hydrophone Based On Parameter Optimization Variational Mode Decomposition And Wavelet Denoising

MEMS vector hydrophone has been widely used in underwater acoustic engineering due to its advantages of vectoriness,small volume,high sensitivity and mass production.However,many different noises and interferences will inevitably be mixed in the underwater data acquisition,which will distort the signal and lead to the lack of practicability of the obtained signal.In order to further detect,identify and Orient the signal,it is very important to eliminate the influence of noise interference.In this paper,the denoising method of MEMS vector hydrophone signal is studied by using the method of variational mode decomposition combined with wavelet threshold processing with parameter optimization.The main research contents are as follows:(1)A wavelet threshold denoising method based on the combination of sine and cosine algorithm(SCA)and particle swarm optimization algorithm(PSO)is proposed to optimize VMD algorithm based on SCA-PSO.This method combines sine and cosine optimization algorithm with particle swarm optimization algorithm to optimize the parameters k and ?of VMD decomposition method,so that it can find the optimal parameter pair k),(?adaptively,and then carry out wavelet threshold denoising.The simulation experiment and the experimental data of Fen machine in North University of China show that the proposed method can effectively remove the noise contained in the source signal under different decibel noise,and it is more superior than the other three algorithms.(2)A wavelet threshold denoising method based on the hybrid algorithm MVO and particle swarm optimization algorithm(PSO)is proposed.According to the theory of cross-relation number,the IMFS components obtained from variational mode decomposition were optimized by MVO-PSO.The IMFS components were divided into three parts: noisy IMFS components,noisy IMFS components and pure IMFS components.The pure IMFS component was retained and the noisy IMFS component was abandoned.The noisy IMFS component was denoised by wavelet threshold.Finally,the denoised noisy IMFS component and the pure IMFS component were reconstructed to obtain the overall denoising signal.The simulation experiment and the experimental data of fen machine in north university of China show that,under different decibel noise,compared with other algorithms,this method is better than other algorithms in both visual denoising effect and qualitative performance evaluation index,and can effectively remove the noise contained in the signal and has certain advantages.In this paper,two new signal denoising methods are proposed.The algorithm has strong practical value and can effectively remove noise,which provides a new idea for denoising vector hydrophone and lays a certain foundation for estimating the direction of arrival Angle of vector hydrophone array in the next step.