Study On Fault Diagnosis Of Rolling Element Bearings Based On Regularized Orthogonal Matching Pursuit Algorithm

Rolling bearings are called mechanical joints as supporting moving parts,which directly affect the operation stability of the equipment and the safety of the working structure.Therefore,the condition monitoring of the bearing operation is of great significance to the normal operation of the machine.The changing trend of the equipment’s operating status is directly reflected by the status data.With the upgrading of production technology,the amount of data processing has increased geometrically.The current important problem in the field of fault diagnosis is the processing and effective analysis of massive data.This research studies the fault diagnosis method of rolling bearing based on the regularized orthogonal matching pursuit algorithm.The compressed sensing theory is used to reconstruct the signal and reduce the noise preprocessing,and a good signal reconstruction result is obtained.In addition,a comparative study was carried out on the feature extraction methods of the fault signal,and the method with better feature extraction effect can be used to effectively identify the feature frequency of the fault signal of the rolling bearing.The main work content is as follows:(1)The research status of fault diagnosis and compressed sensing theory is studied,the structure of rolling bearing and fault diagnosis technology are described,the calculation formula of fault characteristic frequency is given according to the vibration fault mechanism,and the research progress of the method of extracting fault characteristics of rolling bearing is systematically summarized.Lay a good theoretical foundation for the practical application of compressed sensing theory combined with fault diagnosis technology in the following.(2)In-depth research on the theoretical framework of compressed sensing.Select the discrete cosine transform base to sparse the signal,and use the random Gaussian matrix that meets the constraint equidistant characteristics as the measurement matrix;for the selection of the reconstruction algorithm,the five algorithms are respectively reconstructed and analyzed for the simulation signal and the experimental signal,and the running time is compared.Reconstructed signal residuals and reconstructed signal time domain kurtosis and other performance parameters,using better performance regularized orthogonal matching pursuit algorithm for signal reconstruction and noise reduction preprocessing.(3)Theoretical analysis of empirical mode decomposition and its improvement methods,using empirical mode decomposition,ensemble empirical mode decomposition,and complementary ensemble empirical mode decomposition to process analog signals and experimental signals to obtain fault characteristic frequencies,and compare them.The solution effect of the modal aliasing problem and the identification efficiency of the fault frequency and frequency multiplication prove that the complementary set empirical mode decomposition is more effective.(4)Use the fault vibration signals of the inner and outer rings of the rolling bearing to verify the signal reconstruction method and the fault feature extraction method proposed in this research.First,use the compressed sensing theoretical framework of this research to reconstruct the bearing fault signal;then use the complementary set empirical mode decomposition method to decompose the reconstructed signal and reconstruct the components to achieve signal noise reduction.The method in this research reduces the error of fault frequency identification,increases the frequency multiplication of fault identification frequency,effectively eliminates noise,and achieves the purpose of fault diagnosis,which shows the superiority of this method.