Improved Synchrosqueezing Transform Method And Its Application In Rolling Bearing Fault Diagnosis

Mechanical equipment often runs in complex environments,which inevitably leads to performance degradation and failures.Therefore,it is of great significance to carry out the research on mechanical fault diagnosis in complex environments to ensure the safe operation of mechanical equipment.The research on fault diagnosis of key components of mechanical equipment under time-varying speed conditions has become a research hotspot.To realize the condition monitoring and fault diagnosis of mechanical equipment in complex environments,one of the key problems is how to effectively extract and analyze the non-stationary characteristics of mechanical dynamic signals.As a new time-frequency analysis method,synchrosqueezing transform has good time-frequency aggregation and reconstruction characteristics,and is not interfered by cross terms.Therefore,it has been widely studied and applied for the feature extraction of non-stationary signals of mechanical equipment.In this paper,aiming at the problems faced by the existing synchrosqueezing transform in high-resolution time-frequency representation,fault-sensitive frequency band location and fault instantaneous feature extraction under time-varying speed conditions,the vibration signal model of rolling bearing under time-varying speed condition,improved synchrosqueezing transform algorithm,fault sensitive band location and instantaneous fault feature extraction method are studied from three aspects of theory,algorithm and application.(1)The basic structure and motion characteristics of rolling bearings are studied and analyzed.The fault vibration mechanism of rolling bearings is summarized and described,and the fault vibration signal model of rolling bearing to damage type is constructed.It is extended to the condition of variable speed.The calculation formula for fault characteristic frequency of rolling bearing surface damages type is summarized,which lays theoretical support for the simulation of subsequent rolling bearing fault simulation signal and the in-depth analysis of measured bearing fault signal.(2)The fixed window of the traditional time-frequency analysis method has the problems of low time-frequency aggregation and unable to realize the high-resolution expression of time-frequency in the analysis of nonlinear frequency modulation signal.In this paper,the synchrosqueezing transform theory is introduced on the basis of short-time Fourier transform,and the synchrosqueezing transform algorithm of window expansion optimization is proposed by using the local information characteristics of the signal.On this basis,the second-order and high-order algorithms are deduced.The new method can further embody the advantages of synchrosqueezing transform,and further sharpen the time-frequency ridge,so as to enhance the energy aggregation level of time-frequency representation,improve the time-frequency resolution of the signal,and realize the high-resolution expression of the time-frequency domain characteristics of the vibration signals rapid extraction of speed information about the condition of time-varying speed.The experimental simulation verify the effectiveness and reliability of the proposed method.(3)Aiming at the problem that it is difficult to separate the damage fault characteristics of rolling bearings on time-varying speed conditions under strong background noise,a sensitive band location method based on spectral correlation kurtosis map in frequency domain is proposed.This method introduces the spectral correlation kurtosis index in frequency domain instead of kurtosis index into the fast spectral kurtosis graph algorithm and quickly locates the frequency range of common fault signals to rolling bearings by optimizing the spectral correlation kurtosis value.The simulation analysis shows that the proposed algorithm effectively reduces the interference with background noise and other non-fault shock information,and realizes the fast and accurate positioning of the sensitive frequency band of vibration signal under the condition of time-varying speed.(4)In view of the unknown prior knowledge of rolling bearing vibration signal under time-varying speed conditions,the minimum information entropy criterion is used to expand and optimize the intercepted signal window,and the optimal expansion window parameters are adaptively determined.The adaptive window expansion optimization is realized,and the local characteristics of time-varying signals are better characterized.At the same time,in order to highlight the time-frequency ridge information components,an improved dynamic path optimization ridge detection method is proposed to extract the instantaneous frequency ridge curve.The ridge detection method of existing dynamic path optimization algorithm is improved to extract instantaneous frequency quickly.The improved adaptive window expansion synchrosqueezing transform algorithm is applied to the fault diagnosis of rolling bearings on time-varying speed conditions.The effectiveness of the proposed method is verified by multiple sets of experimental data onto different rotational speeds.The analysis results of simulation signals and measured experimental signals verify the effectiveness and practicability of the improved algorithm compared with the existing similar methods,and the ideal application is realized in the fault characteristic frequency extraction of variable speed rolling bearings,which is of great significance of the analysis of instantaneous frequency of variable speed conditions.