Technology Research, The Main Gear Box Fault Diagnosis Based On Wavelet Neural Network In The Labview Car

The reliability of automobile transmission system is the key which guarantees the vehicle power performance, driving performance and economic fuel performance. The driving axle is the most important part but works under the worst condition in the transmission device, especially the gears of final drive. The meshing conditions of driving and driven gears influence transmission performance. Therefore, a method for gear fault diagnosis of the final drive with wavelet and neural network based on LabVIEW is applied to identify quality of the final drive online, in order to avoid reject of retarder, reduce noise and improve the transmission performance of final drive in this paper.Based on introduction about the structure and functions of the final drive, the simplify vibration model is built; gear fault types and reasons of final drive is analyzed; the fault vibration mechanism, the relation of noise and vibration and vibration signal characteristic of different faults are discussed in this paper. The paper designs the overall plan of vibration measurement machine, puts emphasis upon the design of electric system. The hardware system composes of PLC, acceleration sensor, data acquisition card and industry computer. The software system is designed with Lab VIEW, including functions of signal and image acquisition, data analysis and processing, data display and storage, fault diagnosis based on neural network and so on.According to comparison and study, the paper applies the kurtosis coefficient and wavelet packet decomposition method to extract fault characteristic based on experience,information and a lot of experiment data. The vibration signals of gears collected by acquisition system are preprocessed through wavelet transform to decompose into eight vectors of different frequency bands. The energy vectors of six optimal features are used as inputs to the artificial neural network (ANN) in the diagnosis system. The ANN is trained using back-propagation (BP) algorithm with a subset of the experimental data for known gear conditions. The ANN is tested using the remaining set of data. And then the network trained can be used to diagnose gear fault.The vibration measurement machine designed in this paper can be used as fault diagnosis on-line consequently then it supplies judgment standard for product quality control. It is also in favor of reducing noise, improving productivity and reducing cost. This system can be used not only for gear fault diagnosis of final drive, but also for singal acquisition and analysis of other mechanical vibration.