Research On Adaptive Weighted Sparse Diagnosis Method For Aero-engine Bearing

Aero-engine work under extreme conditions such as high speed,high temperature,high pressure,heavy load and strong disturbance for a long time.Performance degradation and failure of the main shaft bearing inevitably occur,resulting in increased maintenance costs and possible accidents.Therefore,it is of great significance to cry out fault diagnosis of aero-engine bearing to ensure the safe and reliable operation of aero-engine.Aiming at the problem that the fault characteristics of aero-engine bearing are disturbed by time-varying noise,an adaptive weighted sparse denoising model is proposed.Firstly,the time-varying characteristics of noise signal variance are studied,and an adaptive weight matrix matching the time-varying structure of noise and the singular value distribution pattern of fault feature is constructed.The bridge between the physical prior of time-varying noise and the energy of sparse coefficients is established,which solves the problem of consistent shrinkage of sparse coefficients in traditional sparse representation algorithms.Then,a block operator matching the feature period prior is constructed,and an adaptive singular value decomposition dictionary of the aliasing variation feature is established to solve the problem that the traditional sparse dictionary is difficult to match the aliasing variation fault waveform.Finally,an adaptive weighted sparse denoising algorithm is developed based on the block coordinate optimization framework,and the algorithm parameter selection method is studied through numerical experiments.Simulation analysis and aero-engine bearing experiments verify the noise reduction performance of the algorithm.Aiming at the problem of aliasing variation characteristics and strong harmonic signal coupling of aero-engine bearing,an adaptive double-weighted sparse decoupling diagnosis model is constructed.Firstly,based on the time-varying characteristics,an adaptive noise weight matrix is constructed to effectively characterize the time-varying noise interference of aero-engine bearing.Then,based on the periodic modulation intensity index,an adaptive singular value weight matrix is constructed to solve the coupling problem of aliasing variation fault waveform and harmonic interference.Finally,based on the alternating direction multiplier algorithm theory,an adaptive double-weighted sparse decoupling algorithm is developed,and the regular parameter selection method is studied through numerical experiments.Simulation analysis and aero-engine bearing experiments verify the decoupling effect of the algorithm.Aiming at the problem that the fault source signal of aero-engine bearing is disturbed by transmission path modulation and time-varying noise,the adaptive noise weight matrix is introduced into the convolution sparse learning model,and an adaptive weighted sparse deconvolution model is constructed.Based on the multi-block non-convex alternating direction multiplier method framework,an adaptive weighted sparse deconvolution algorithm is developed.Simulation analysis and aero-engine bearing experiments verify the unwinding effect and diagnostic effectiveness of the algorithm.Finally,an adaptive weighted sparse diagnosis system for aero-engine bearing is developed by integrating the sparse diagnosis algorithm proposed in this paper.The overall framework of the system is designed,and the graphical user interface is developed by using the App Designer toolbox of MATLAB.The system can quickly and easily diagnose the fault of aero-engine bearing.