Rolling Bearing Fault Diagnosis Under Variable Speed Conditions Based On Synchrosqueezing Transform

Rolling bearings are widely used as key components of rotating machinery and equipment,whose operating conditions are directly related to the entire mechanical system.Rolling bearings are subject to variable operating conditions such as variable speed and load,and are prone to break down.Therefore,it is of great significance to study how to monitor and diagnose the condition of rolling bearings under variable speed conditions.This subject is based on the synchronous compression transform(Synchronous Compression Transform(SST))and uses the characteristics of rolling bearing fault signals under variable speed conditions to make a series of improvements to SST to achieve fault diagnosis under different conditions.The main work of this subject includes:Firstly,elaborating the background and significance of the topic systematically based on the monitoring of the operating conditions of industrial rotating equipment.This paper comprehensively discusses the analysis methods of rolling bearing vibration signals under variable speed conditions,the structural form of rolling bearings,the types of failures and the vibration mechanism;and introduces the calculation method of fault characteristic coefficients for different parts of variable speed rolling bearings.Secondly,in order to solve the problem that SST cannot achieve optimality in the selection of window length,two criteria,Rényi entropy criterion,which is selected through experimental methods,and maximum kurtosis criterion,are proposed for optimization.According to the characteristics of vibration signals of rolling bearings under the conditions of lifting and lowering speed and the optimal window length SST,a time-frequency demodulation spectrum scheme is proposed to solve the problem of extracting the instantaneous fault characteristic frequency of rolling bearings under the condition of lifting and lowering speed,and then compared with the theoretical value of IFCF for different faults to achieve fault diagnosis.Through simulation signals and measured data,the optimal window length SST can be used to extract and diagnose IFCF of rolling bearings,which verifies the effectiveness of the scheme.Thirdly,under the condition of no tachometer,the low-frequency vibration signal of the rolling bearing is used to realize the instantaneous frequency estimation.However,the signals are complicated,making the resolution of the time spectrum low and blurring,which will directly affect the accuracy of the instant frequency estimation.A transient frequency estimation algorithm based on Variational Mode Decomposition(VMD)-SST is proposed: First,the vibration low-frequency signal is VMD-decomposed,and the component signal is selected for reconstruction according to the correlation coefficient.Second,the reconstructed signal is subjected to SST processing to obtain the time spectrum.Finally,the Viterbi algorithm is used to estimate the instantaneous frequency conversion in the time spectrum.Simulation experiments and measured data verify that the scheme can not only improve the noise immunity of SST and the resolution of the frequency spectrum,but also improve the accuracy of the instantaneous frequency estimation.Combined with the algorithm of extracting IFCF in Chapter 3,the fault diagnosis of rolling bearing under tachometer condition is realized,which has certain application value.Finally,under the condition of large fluctuation speed,the frequency spectrum of the rolling bearing vibration signal is prone to energy divergence at the fluctuation,which makes it difficult to extract the characteristic frequency of the turning fault.Multiple compression transforms can change this situation,but it is difficult to choose the number of compressions.Therefore,fault feature extraction based on adaptive multiple compression transform is proposed.In this chapter,Rényi's entropy order spectrum is introduced to realize adaptive multiple compression transforms.The simulation signals verify the advantages of the algorithm in this chapter in processing emphasized frequency signals and multi-component cross interference signals;and the method is successfully applied to extract the characteristic frequency of rolling bearing faults in large fluctuation speed conditions.