Matching Synchroextracting Transform for Mechanical Fault Diagnosis Under Variable-Speed Conditions
Yubo Ma, Yong Lv, Rui Yuan, Gangbing Song
Abstract
Time–frequency (TF) analysis (TFA) technique has been widely used to the analysis of rotating machine vibration. However, vibration signal from practical sources contains complicated components and noise, so the fault diagnosis to variable-speed machinery is full of challenges. In this study, inspired by the demodulated synchrosqueezing transform (DSST), the synchroextracting transform (SET) is extended to a prior instantaneous frequency (IF) based method, named matching synchroextracting transform (MSET). To achieve the fault diagnosis using MSET, the follow-up works mainly include two parts. First, a demodulation filtering strategy is developed for multicomponent signal separation. Second, the order analysis based multiple IFs estimation idea and the second-order difference operator based IF smoothing scheme are introduced to obtain the reliable initial IFs. The effectiveness of the proposed technique is verified via some simulation studies. Finally, the proposed technique is successfully applied to the fault diagnosis of rolling bearing and planetary gearbox.