| Mechanical equipment usually works in harsh environments,such as high temperature,high speed,etc.These factors often lead to performance degradation or even failure of the equipment.The failure of key mechanical equipment will bring great harm to production and operation and work safety.In order to ensure the long-term stable operation of mechanical equipment,it is necessary to regularly diagnose the operating status of the equipment.In practical engineering applications,the vibration signal that can reflect the operating state of mechanical equipment often exhibits strong non-stationary characteristics.Aiming at the problems that the traditional time-frequency analysis method is difficult to deal with strong frequency signal,poor time-frequency energy concentration and inaccurate pulse mode location,three improvement schemes are proposed based on synchrosqueezing S-transform in the paper.(1)Demodulated synchrosqueezing S-transform(DSSST),which is used to characterize the oscillation characteristics of strong frequency modulated signal,(2)Multisynchrosqueezing S-transform(MSSST),which improves the energy concentration of timefrequency representation in the frequency direction through multiple iterations to instantaneous frequency operator,(3)Time-reassigned multi-synchrosqueezing S-transform(TMSSST),which enhances the energy concentration of time-frequency representation in the time direction through multiple iterations to group delay operator.The specific content is as follows:(1)In order to cope with strong frequency modulated signal,a demodulation algorithm is utilized to process the vibration for reducing the non-stationarity,and we propose DSSST.This method can extract the oscillation feature of several modes and improve the energy concentration of time-frequency representation.The theoretical frame of DSSST is constructed,and the algorithm implementation is given.Multi-component strong frequency modulated signal,experimental signal and engineering signal verify the effectiveness of DSSST in rotor rubbing fault diagnosis.(2)In order to deal with multi-component signal,the MSSST is proposed by multiple iterations of the instantaneous frequency operator.The instantaneous frequency operator after multiple iterations is very close to the real instantaneous frequency,which greatly improves the energy concentration of time-frequency representation in the frequency direction.The theoretical framework of MSSST is built,and the algorithm implementation is given.The gear fault simulated signal demonstrates the advantage of MSSST in time-frequency energy concentration and noise robustness,and its engineering application.(3)In order to handle the impulse signal,the TMSSST and the noise reduction reconstruction algorithm are proposed based on multiple iterative group delay operator and impulse feature retrieval technique.This method can greatly improve the energy concentration of time-frequency representation in the time direction,and filter out random noise.The theoretical framework of TMSSST and its reconstruction expression are constructed.Bearing fault signals under various working conditions illustrate the effectiveness of TMSSST in bearing fault diagnosis. |
