采用时序分析和BP神经网络,建立了基于时序-神经网络的车辆变速器齿轮故障诊断系统。通过

对车辆变速器齿轮运行状态特征信号进行时序分析和特征向量提取,并以此作为BP神经网络的输入向量进行网

络训练,从而实现变速器齿轮运行状态的识别与故障诊断。该系统应用于LC5T81变速器齿轮的故障诊断中,能

够比较准确地识别与诊断出变速器齿轮的跑合运行状态、磨损运行状态和故障运行状态。验证表明该诊断系统有

效、可行。

-Fault Diagnosis of Vehicle Transmission Gear Based on Time

Series Analysis and Neural Networks

Yin Andong & Yang Zhengmin

School of Mechanical and Automobile Engineering,Hefei University of Technology,Hefei230009

　　[Abstract]　Based on time series analysis and BP neural networks, a fault diagnosis system is built for ve-

hicle transmission gears. By time series analysis and eigenvectors extraction on operation status signals of trans-

mission gears, which are taken as inputs for neural network training, the operation status identification and fault

diagnosis for transmission gears are realized. In a fault diagnosis on the gears of a real transmission, the system

can accurately identify the operation status (running in, worn or fault). The result shows that the system is ef-

fective and feasible.