Application Of Generalized Demodulation Time-Frequency Analysis Method In Fault Diagnosis Of Rolling Bearings

Bearings are the key components in industrial production equipment,which are prone to failure due to the high rotating speed and heavy loads for long-term running.Meanwhile,with the fast development of rotating machineryThe fluctuations of speed wave should not be overlooked.Vibration signals during the process of speed changing contains rich information of bearing faults and has unique value to fault diagnosis.Therefore,fault diagnosis method of rolling bearings under variable speed condition is researched in this paper.A tacholess order tracking method is proposed whose key part is the acquiration of the rotating speed signal based on instantaneous frequency estimation.The generalized demodulation time-frequency analysis method(GMTA)and its application in the estimation of instantaneous frequency are fully studied.The main research work and achievements are as follows:First of all,considering that vibration signals under variable speed working conditions present the characteristics of non-stationary,multi-components and time-varying,a tacholess order tracking propose based on fast spectral kurtosis is presented,which could effectively remove the influence of speed changing without any extra hardware.It can well diagnose and analyze the fault impulse signals.Secondly,an estimation approach of instantaneous frequency based on the GMTA method is researched to achieve the rotating speed signals.Through the selection of appropriate phase function,the GMTA method can be utilized to analyze the fault vibration signals well and effectively.Simulation analysis shows that the proposed GMTA method is suitable to processing such multi component AM/FM signals like bearing fault vibration signals under variable working conditions.Moreover,compared to the EMD method,it can reduce the end effect as well as improve the estimation accuracy of instantaneous frequency.Finally,the proposed tacholess order tracking method is performed on the bearing simulation signals with changing speed condition.Firstly,the fast kurtosis method is utilized to extract the fault impulse signals.Then it is resampled in angle domain to remove the influence of the speed variation.The results show the feasibility and effectiveness of the presented method.