Tractoris a complex device in the basic components of the whole machine, which is composed of gear pair, shaft, bearing, box, lubricating oil, etc. During ploughing, rotary tillage and transportation of tractor, the unit is mainly affected by the torque of driving wheel, cutter shaft resistance torque and traction. Most of the faults occur in mechanical transmission systems such as gearbox and transmission shaft. The incidence rate of gearbox fault is 50% 1 70% of other faults except engine failure.
。 Therefore, it is of great significance to study and explore the mechanism, mode and diagnosis method of tractor gearbox fault for the use and maintenance of the whole tractor.
In the previous analysis and processing of gearbox fault vibration signal, it is usually considered that the measured signal is a stationary signal, so Fourier analysis is widely used in signal analysis and processing. However, in practical work, due to the impact caused by unstable machine speed, load change and machine failure, friction will lead to the reduction of non-stationary vibration signal
Using Fourier analysis to analyze these abnormal signals may get false results or diagnosis results, which greatly affects the accuracy of diagnosis results. Wavelet analysis has incomparable advantages in dealing with the above problems.
It can decompose any signal stationary or nonstationary into a basis function family formed by the expansion of wavelet basis. The amount of information is complete. The decomposition sequences distributed in different frequency bands are obtained in the pass frequency range. It has the localization analysis function in both time domain and frequency domain. Especially when the early weak defects hidden in a part of the machine, its defect information is submerged by the vibration signals and noise of other parts. When it is difficult to obtain satisfactory results by time-domain method and Fourier analysis method, continuous wavelet transform can better solve the problem of accurate diagnosis of this kind of gear box.
In addition, wavelet packet analysis also has a wide application prospect in gearbox fault diagnosis. It can separate the useful components, interference and noise in the original signal, and has a certain adaptability to frequency drift. These characteristics determine that wavelet packet analysis is very suitable for combining with artificial neural network to construct gearbox automatic fault diagnosis system.
On the basis of wavelet analysis of fault vibration signals, combined with artificial neural network technology, this paper discusses how to establish an automatic fault diagnosis system for tractor gearbox, in order to expect a breakthrough in tractor gearbox fault diagnosis.