Research On Local Fault Diagnosis Method Of Rotating Machinery Based On Robust Sensing Matrix

With the development of science and technology and the advancement of industrial modernization,the structure of rotating mechanical equipment becomes more and more complex.The real-time fault monitoring of mechanical equipment is an important method to ensure the stable operation of the whole mechanical system and prevent production accidents.The real-time fault diagnosis of mechanical equipment is faced with the problem of compression,transmission and storage of massive real-time data.How to compress and reconstruct the collected signals more efficiently is a major challenge for mechanical fault diagnosis system.Based on compressed sensing theory,this paper analyzes and studies the optimal design method of robust sensing matrix and its application in fault diagnosis of rotating machinery,which provides a theoretical basis for solving the above problems.Aiming at the problem of poor performance of the sensing matrix used in the current mechanical fault diagnosis system,the optimization design method of sensing matrix is introduced into the mechanical equipment fault diagnosis system,which can achieve a greater degree of compression of the collected data on the premise of ensuring the effective reconstruction of the signal.Considering the characteristic of low signal-to-noise ratio of mechanical vibration signals,a robust sensing matrix optimization design framework suitable for mechanical vibration signals is obtained based on analyzing the anti-noise performance of different sensing matrix optimization design frameworks.Based on the optimization design framework,an analytical solution algorithm with lower computational complexity is derived,which can avoid singular value decomposition and improve the speed of solving the sensing matrix.Aiming at the problem that the dense sensing matrix is not conducive to hardware implementation,the optimal sensing matrix is sparsified and the influence of different sparsification strategies on the reconstruction accuracy of the compressed sensing system is studied to achieve a good balance between the performance of compressed sensing and the difficulty of hardware implementation.To solve the problem of low reconstruction accuracy of mechanical vibration signal under strong noise,the design method of improving the robustness of sensing matrix is studied.The mechanism of the regularization term used to improve the robustness of the existing sensing matrix design framework is analyzed,and the limitations of the existing robust sensing matrix design methods are analyzed based on the mechanism.Based on the mechanism of the robust regularization term,a new robust sensing matrix design method is proposed,which overcomes the limitations of the existing methods and has better anti-noise performance.The simulation and experimental results verify the analysis of the mechanism of the robustness regularization term.Experimental results on fault bearings and fault gearboxes show that the proposed method not only has high solving efficiency,but also can deal with mechanical signals with low SNR,which is beneficial to identify weak fault features.The compressed sensing system based on the proposed method can effectively reconstruct the mechanical fault signal at a lower compression rate and has better compressed sensing performance.