Gear Faults Diagnosis Based On Wavelet Neural Network

With the development of the science and technology, it becomes the mechanical equipments' tendency that people seek for the high performance, the high efficiency, the high automation and the high reliability. Because of the advantages of the fixed transmission ratio, the larger transmission torque and the compact structure, gear is widely applied to adjust velocity and transfer power. Because gear is an important component of the most machine equipments, it arises some faults frequently, which leads to the poor performance of the whole machine in the operating condition. Therefore, more and more researchers throw their attention on the fault diagnosis of the gear case's research. And also a lot of research results have been presented since the last two decades.Aiming at the actual gear fault and hard monitoring's problem, firstly, the dynamics theories is made use of to establish wheel gear vibration system model in this thesis. Through the theories of analysizing a few familiar typical gear fault model, we get the influence of the gear vibration signal. Secondly, WNN is more superior of study speed, astringency, association of thought than the BP network by making use of imitating tests. Especially for the scale bigger network, the advantage would be more outstanding. Thirdly, the WNN model suitable for wheel gear is set up to diagnosis gear faults, the WNN studying calculation adopts the BP calculation of self-adjusting the studying velocity. The characteristic parameters of the gear fault vibration signal are passed to train WNN, adjusting the WNN's power parameters, the gear faults are identified. Finally, an expert system of gear faults diagnosis based on the Wavelet Neural Network is constructed by using of the Visual Basic programe. By analysising the imitating gear faults' signal, we could know that the intelligence diagnosis of the gear fault is realized, which lays the foundation for application in the spot diagnosis.The result indicates the effectiveness of this method because it can improve the efficiency of gear diagnosis and also have a good future.