Research On Fault Diagnosis Of Rolling Bearing Based On Sparse Representation Principle

With the development of science and technology,the mechanical equipment becomes larger,higher speed and automation.Its running state is related to the development of national economy and the safety of people’s life.As a key part of mechanical equipment,rolling bearing will directly or indirectly affect the normal operation of the equipment.Therefore,fault diagnosis of rolling bearing is of great significance.Because the working environment of rolling bearing is complex,the vibration signal is non-stationary and nonlinear complex signal,which makes it difficult to analyze the fault signal.As a widely used rolling bearing fault diagnosis method in recent years,sparse representation can represent signals with less atoms,thus reducing the complexity and redundancy of bearing signals and improving the efficiency and accuracy of fault diagnosis.This paper focuses on the fault diagnosis method of rolling bearing based on sparse representation principle.The specific contents are as follows:Firstly,in view of the problem that the pulse characteristics in early weak fault diagnosis are easily submerged by noise signals,based on the coherence between signals,an autocorrelation function pulse energy ratio(Autocorrelation Function Pulse Energy ratio,ACFPER)is proposed to identify the components of the signal that can represent the feature information.And apply it to the sparse representation of the signal.The feasibility and effectiveness of the index are verified by data simulation.Then,a G-KSVD dictionary learning method based on effective singular components is proposed to solve the problem of missing optimal atomic selection in traditional learning dictionaries.ACFPER is used as the index to filter the singular components,to reduce the noise of the signal,and to update the dictionary atoms by using the components containing more fault information,so as to enhance the impact components in the signal.The time cost is reduced by reducing the feedback layer.Simulation and actual signal experiments show that the G-KSVD algorithm has denoising effect and low time cost in the effective interval.Finally,AAFBP(Adaptive Acceleration Forward-Backward Pursuit)algorithm is used to reconstruct the signal,envelope demodulation is performed on the reconstructed signal,and the signal is analyzed by its characteristic frequency.The experimental results show that this method can accurately identify fault information,improve the accuracy of rolling bearing fault diagnosis,and promote the development of fault diagnosis technology.