A Parameter-Optimized Variational Mode Decomposition Investigation for Fault Feature Extraction of Rolling Element Bearings

Reliable fault diagnosis of the rolling element bearings highly relies on correct extraction fault-related features from vibration signals in time-frequency analysis. However, considering nonlinear, nonstationary characteristics signals, hidden heavy noise has become a challenging task. Variable mode decomposition (VMD) is an adaptive, completely nonrecursive method variation and signal processing. This paper analyzes advantages VMD compared with EMD robustness against noise, overcoming end effect aliasing. The performance algorithm largely depends selection number k bandwidth control parameter α. To realize adaptability influence parameters improvement accuracy, parameter-optimized presented. random frog leaping (SFLA) used to search optimal combination parameters, are set according results. A multiobjective evaluation function constructed select component. envelope spectrum technique analyze proposed evaluated by simulation practical bearing under different conditions. results show that can improve accuracy effective signal.