An Improved VMD With Empirical Mode Decomposition and Its Application in Incipient Fault Detection of Rolling Bearing

Abstract

Transient impulse analysis is an effective way to detect the bearing fault at its early stage. However, it is hard to precisely extract these so-called transient impulses because these collected vibration signals usually are non-stationary, nonlinear, and drowned by heavy background noise. Variational mode decomposition (VMD) can play the role as an adaptive signal processing tool to reveal the weak transient impulses from complex vibration signals. However, its reasonable mode number is difficult to pre-set and this would make the loss of useful transient impulses. To solve this issue, an improved VMD strategy is presented in this paper. For this method, it can not only utilize the advantages of traditional VMD and empirical mode decomposition (EMD) but also adaptively select sensitive intrinsic mode function (IMF) components for fault component analysis by proposed indexed values. EMD is first used to process the collected vibration signal into a series of IMFs, and the so-called useful IMFs are then evaluated by a sensitive IMF evaluation index which is based on the conjoint analysis of relatedness and kurtosis. Afterward, VMD is further improved to effectively decompose the denoised signal reconstructed from these selected useful IMFs of traditional EMD. Finally, the improved VMD is used for incipient fault diagnosis by a defined transient impulse monitoring index and Hilbert envelope analysis. Experiments are performed to demonstrate the effectiveness of the proposed method. The experimental results confirm that the proposed method can accurately extract the features of an incipient fault of a bearing.