Identify The Faint Defect Of Bearing Based On Model

With the development of automobile industry and appliance, it requires the mechanical parts with high precision. As the main supporting part, the vibration performance of rolling bearing directly affects the whole equipment. The current bearing diagnosis technique focus on testing the invalidation of working bearing. Because of the errors during the manufacture , it is hard to distinguish the fault signal of new bearing from other signals. This article wants to find out an available method based on model to detect the bearing fault rapidly and effectively. The factory can improve their craft timely with this method in order to decrease the noise.In this article, the wavelet packet analysis method is introduced and the eigenvector of bearing state is found through the study of bearing fault signal. On the other hand, the vibration model based on the tribo-dynamics viewpoint is constructed ,which is the theory basis of signal analysis. The results are as follows:1,A kind of new wavelet is constructed to deal with the fault signal.2,The wavelet packet combines with the auto—regression spectrum to distinguish the bearing signal.3,Study the bearing signals and find the position of frequency which the defect occurs from the different frequency band and suggest the vibration frequency mode function of normal bearing .4,The differential equation of bearing is established overcoming the linearization hypothesis of classic bearing vibration model, which considers bearing non-linear dynamics characteristics. The defect trait of waviness in balls and rings is discussed and calculated. And then the method based on model combines with the signal analysis to identify the faint defect.