EWT-fused ISSA-VMD-KU uwDAS vibration signal denoising method

The invention discloses a uwDAS vibration signal denoising method of ISSA-VMD-KU fused with EWT. The method comprises the following steps: collecting a vibration signal y (t) by using uwDAS; performing EWT preprocessing on the vibration signal y (t) to obtain M mra modal components; calculating a Pearson correlation coefficient rho of each order mra and the original signal, and selecting a component in a threshold range rho = [rho l, rho h] for reconstruction to obtain a reconstructed vibration signal y '(t); using VMD to decompose the reconstructed vibration signal y '(t), using the minimum envelope entropy as a fitness function, using ISSA to optimize the decomposition layer number K and the penalty factor alpha of the VMD until the maximum number of iterations is reached, and outputting an optimal parameter combination [Kbest, alpha]; the optimal parameter [Kbest, alpha] is used to carry out VMD decomposition on the y '(t), and Kbest IMF and a remainder RES are obtained; and calculating a kurtosis value KU