A coarse-to-fine decomposing strategy of VMD for extraction of weak repetitive transients in fault diagnosis of rotating machines

Variational Mode Decomposition (VMD) has attracted much attention and been used to analyze different kinds of signals, such as mechanical signals, medical data, and financial time series, etc. However, the VMD is still confronted with some dilemmas during the applications, including the determination of the number of the decomposed modes, the selection of the balance parameter, and so on. To address these problems of the VMD, a coarse-to-fine decomposing strategy is proposed for weak fault detection of rotating machines in this paper. Firstly, through extensive numerical simulations, the characteristics of the relative bandwidths of the decomposed modes are given with the change of the balance parameter and the number of the decomposed modes. Then, motivated by the bandwidth characteristics, the rationalities and advantages of iterative decomposition of the VMD and the fine adjustment of the balance parameter are discussed in detail, respectively. Subsequently, the new coarse-to-fine decomposing strategy of the VMD is developed to obtain the optimal mode and extract the weak repetitive transients of rotating machines. The analysis results of the simulated signals and the experimental signals measured from two run-to-failure cases show that the proposed method can well-detect the weak repetitive transients in the signals with heavy noise and overcome the drawbacks of the original VMD. The superiority of the proposed method for faint repetitive transient detection is also demonstrated by comparing with the existing methods.