Fault feature extraction of rolling element bearing based on TPE-EVMD

Abstract The periodic impact components generated by bearing defects are often weak and submerged by strong background noise due to the influence of complex transmission path. However, little attention has been drawn to this content. To address this problem, a fault diagnosis method for rolling element bearing was proposed using the transmission path elimination based enhanced variational mode decomposition (TPE-EVMD). First, the relationship among transfer function, exciting and response signal was analyzed, a trend-line based transmission path elimination technique was constructed. Second, spectrum editing method was employed to both ensure the accuracy of trend-line’s calculation and suppress the existence of harmonic components. Finally, a whale optimization algorithm based variational mode decomposition was utilized to further suppress the original noise. The simulation and experimental verification indicate that the proposed technique can extract the fault signatures of rolling element bearing more accurately and effectively under the influence of complex transmission path.