Velocity Synchronous Chirplet Extracting Transform: An Effective Tool for Fault Diagnosis of Variable-Speed Rotational Machinery
Yong Lv, Yubo Ma, Rui Yuan, Site Lv
Abstract
Bearing vibrations from actual sources always involve noise and other interferences, causing the poor time-frequency (TF) readability of time-frequency distribution (TFD) and the inaccurate estimation of instantaneous frequencies (IFs), thus bringing great challenges to the diagnosis of bearing failures. To this end, under the theoretical framework of synchroextracting transform (SET), this work studied a novel time-frequency analysis (TFA) technique named velocity synchronous chirplet extracting transform (VSCET). Compared with other traditional TFA techniques, the VSCET method can accurately depict the variation laws of time-varying features via a highly readable and energy-concentrated TFD. It actually introduced the basis function in velocity synchronous linear chirplet transform (VSLCT) that has the synchronous changing law of rotational speed into SET such that has better advantages in both energy concentration and noise robustness. In addition, a fault diagnosis scheme based on VSCET is further introduced to realize the identification of bearing faults. The validity and practicality of this technology are fully tested from the results of simulations and experiments.