Adaptive Variational Mode Decomposition Method And Its Application In Mechanical Fault Diagnosis

Variational mode decomposition(VMD)is a new decomposition algorithm,which has good adaptability and non-recursiveness.VMD algorithm has strong noise robustness based on its Wiener filtering principle in the process of signal decomposition and has good result in processing non-linear and non-stationary signal.This paper studies the variational mode decomposition optimization algorithm and applies it to mechanical fault diagnosis.The concrete researched content as follow:(1)For impact and modulation characteristics of the mechanical fault signal,the feasibility of VMD for fault mechanical is studied in this paper.By illustrating and analyzing the principle of the VMD algorithm,the different simulation models are built to verify the decomposed effect of VMD.A variational mode decomposition method based on spectral kurtosis(SK)is proposed for the extraction of mechanical equipment fault features The effectiveness of the proposed method is proven by verifying simulation and experiment data of the bearing.At the same time,comparing with the result of EMD,the results further verify the feasibility of the method of VMD based on SK.(2)In VMD algorithm,the numbers of decomposed modes and penalty factor parameter have a good influence on the performance of VMD algorithm decomposition results.An adaptive variational mode decomposition combined with spectral kurtosis based on particle swarm optimization(PSO)is used to adaptively select parameters.The simulation model and experiment data are used for testifying the feasibility and effectiveness the proposed method and the results show a good effect.(3)The combination of adaptive variational mode decomposition and detrended wave analysis is used for fault feature extraction and fault state recognition of gears.After the adaptive VMD method is use to decompose raw signal and select the high-frequency mode component,DFA is employed to extract the gear fault feature vector.The higher recognition accuracy of the gear fault classification demonstrated that the proposed method can effectively identify the fault state of the gear.