Study On Fault Diagnosis Of Rolling Element Bearings Based On Improved HHT And SVM

With the progress of technology and the development of modern industry, rolling bearing as an important rotating part in the mechanical equipment, is widely used in the field of industry, and it’s the source of fault. Many mechanical equipment failure is associated with rolling bearings, whether the rolling bearing working normally or not affects the performance of mechanical equipment directly. Therefore, it has important theoretical value and practical significance to carry out the research on the fault diagnosis of rolling bearing.This thesis set about from the bearing vibration signal and carry out deep research on the method that fault diagnosis involved, the main contents are as follows:The contents and significance of rolling bearing fault diagnosis technology is analyzed, the rolling bearing vibration mechanism and the fault characteristic frequency is introduced, the development status of the rolling bearing fault diagnosis of feature extraction and pattern recognition are reviewed.The basic principle of Hilbert-Huang transform and the concept of Hilbert spectrum and marginal spectrum is introduced, aiming at mode mixing problem of the empirical mode decomposition, the complete empirical mode decomposition method is introduced. The periodical impulse signal is analyzed by empirical mode decomposition and complete empirical mode decomposition, the intrinsic mode function by complete empirical mode decomposition is closer to the ideal value.Feature extraction method based on the wavelet packet and complete empirical mode decomposition is put forward. The method is used for the simulation signal and the experimental data, the results show that the improved HHT method can effectively extract the fault feature frequency components and suppress noise.Using least squares support vector machine model to simulate the rolling bearing pattern recognition, and using particle swarm optimization algorithm to optimize the parameters, the singular value of the intrinsic mode function is used as the feature vectors of support vector machine, the identification of test samples show that the support vector machine is effective.