Instantaneous Frequency Estimation Of Rolling Bearing Under Variable Speed Conditions Based On Synchrosqueezing Transform

Rolling bearing is a key component of mechanical equipment,its working state directly affects the safe and stable operation of the equipment,so fault diagnosis of rolling bearing is very important.Among them,fault diagnosis under variable speed is the focus and difficulty of research,and the instantaneous frequency estimation is of great significance for fault diagnosis under variable speed conditions.Synchrosqueezing transform is an effective time-frequency analysis method to describe the instantaneous frequency characteristics of non-stationary signals,but there are many problems,such as poor noise robustness,low time-frequency resolution,energy divergence when processing multi-component signals and strongly timevarying signals and so on.In order to solve the above problems,this paper studies a series of improved algorithms and applies them to the instantaneous frequency estimation of variable speed rolling bearings.From the aspects of fault modeling,fault characteristic frequency and instantaneous frequency extraction,the scheme design,performance analysis and comparison of algorithms,and experimental verification are carried out.The main work is as follows:Aiming at the problem that SET window length can not be selected accurately,a method of selecting the best window length by using Rényi entropy index is proposed.In order to improve the noise resistance and eliminate the influence of complex sideband,a time-frequency demodulation spectrum scheme based on optimal window SET is proposed to extract the fault characteristic frequency of rolling bearings with variable speed.Firstly,the fast spectral kurtosis is used to find the optimal frequency band of fault sensitive components,and the vibration signal is filtered by band-pass filter.Then the amplitude envelope of the filtered signal is obtained by Hilbert transform.Finally,applying the optimal window SET to the amplitude envelope signal,and the time-frequency demodulation spectrum can be obtained to extract the fault characteristic frequency.The anti-noise and time-frequency resolution performance of the scheme are simulated and analyzed,and the robustness of the filter under fault state is tested.The experiment validates that the fault characteristic frequency curves are successfully extracted by using the fault signal analysis of inner and outer rings.The results show that the scheme can accurately extract the fault characteristic frequencies of variable speed rolling bearings.In order to improve the time-frequency resolution and the accuracy of instantaneous frequency estimation,the method of instantaneous rotational frequency estimation based on VMD-SET is studied.Firstly,the signal is decomposed into variational modes.Then,the effective components are selected according to kurtosis criterion and mutual information,and the reconstructed signal is obtained.The reconstructed signal is analyzed by synchroextracting transform to obtain the timefrequency expression.Finally,the peak search is used to extract the instantaneous rotating frequency from the time-frequency diagram.The multi-component signals of LFM and sinusoidal FM are simulated and analyzed,and the noise resistance and instantaneous frequency estimation accuracy are evaluated.Experiments verify that the instantaneous rotating frequency curves are successfully extracted from the analysis of the two working conditions of speed-up and complex speed-change.The results show that this method is a high precision instantaneous frequency estimation method.In order to improve the instantaneous frequency estimation ability aiming at the strongly time-varying signal,the multi-synchrosqueezing transform is studied.The time-frequency analysis ability of multi-synchrosqueezing transform in processing multi-component time-varying signals disturbed by noise,strongly time-varying signals and cross-interference multi-component signals is studied,and compared with a large number of improved time-frequency analysis methods,the results show that multi-synchrosqueezing transform has obvious advantages in instantaneous frequency estimation of the above signals.Finally,it is successfully applied to extract fault characteristic frequencies of variable speed rolling bearings.This method provides an effective way to extract the instantaneous frequency of complex signals more conveniently.