The theory of the fault diagnosis for automotive transmission has developing for many years. But the vibration signals in automotive transmission are complex, so the identification of the fault patterns will refer to many other theories. And the rolling elements bearing is the accessory in common use and invalidation easily, exceptional in the automotive transmission. The motive of this paper is to find some methods accurately and effectively for the bearing fault diagnosis. Firstly,the test rig was built and the vibration signal could be gather from this device. For researching the bearing fault diagnosis effectively, the models, the elements and the reason of the fault are analyzed particularly. And base on the bearing vibration signals, the signal processing methods, such as FFT, Hilbert and the Spectrum are researched. The effect of these methods are proved by the processing results. Finally, in VC++ and in Matlab, the Expert System for the fault diagnosis of the rolling element bearing was built. The experimental results show that it can identify the fault patterns accurately. At the same time, the Neural network was built in Matlab to validate the effective of the Expert System. The experimental results show the BP Neural network are effectively for the diagnosis.
|