An Optimized Vmd Method Based on Modified Scale-Space Representation And its Application In Rolling Bearing Fault Diagnosis

Variational Mode Decomposition (VMD) is a signal decomposition algorithm with excellent denoising ability. However, the drawback that VMD is unable to determine the input parameters adaptively seriously affects the decomposition results. For this issue, an optimized VMD algorithm based on modified scale-space representation (MSSR-VMD) is proposed. Before processing the signal by VMD, its spectrum is divided into several frequency bands by modified scale-space representation, acquiring modes' number and the initial center frequency for each mode adaptively. Moreover, a pre-decomposition step is added to the original VMD to determine the target mode, and its penalty factor is adjusted during the iterative update of the VMD to achieve accurate extraction for the fault features. MSSR-VMD and other adaptive decomposition algorithms are employed to handle the simulated and experimental signals separately. By comparing the analysis results, the method has certain superiority in rolling bearing fault feature extraction