Artificial Neural Network-based Gear Fault Diagnosis

Worked as the main components of the transmission system, integrated transmission's performance affects the whole capability of the vehicle. Because of the complicated work environment, integrated transmission always goes wrong. For this reason, it is very important to make diagnosis to integrated transmission.This paper is mainly researched into the fault diagnosis and artificial neural network, on base of it, this dissertation discussed the mechanism and the character of integrated transmission system, introduced the time domain and frequency domain analysis of signal applied to the fault diagnosis.This dissertation discussed the choice criterion of the signal parameter, made experimentation with real situation, got the vibration acceleration signal and went deep into analysis. This dissertation built the model of the BP neural network and confirmed the parameters of the network.On the base of VC 6.0, the dissertation developed a diagnose system of the integrated transmission. The validity of the analysis method proposed in this thesis was proved through collected signal and simulated signal. The result indicates the efficacy of diagnosis and the fine future.