摘要
设计了神经网络诊断系统,并用于融合分析废气含量、发动机转速、氧传感器信号及某些工作状态信息诊断汽车故障。设计方法为应用遗传算法的复制、交换、变异过程代替BP网络的反向传播过程,并对遗传算法进行改进研究。实践证明,这种基于遗传神经网络方法的故障诊断系统具有收敛速度快、推广性能强的特点,提高了汽车故障诊断系统的效率和准确性。
A neural network diagllosis system is designed to integrate and analyze contents of exhaust gas,engine speed.oxygen sensor signal and some work status information and then to diagnose automotive failures.Its design method is to substitute the copy,interchange and variatiou processes of the generie algorithm for the reverse propagation process of the BP network.and improvement of the generic algorithm is also studied.It is proved that the generic-NN based failure diagnosis system converges fast and is easy to be used widely,and it improves efficiency and precision of automotive diagnosis system.
出处
《汽车技术》
北大核心
2005年第9期36-38,共3页
Automobile Technology
关键词
排放
神经网络
故障
诊断
Emission,Neural network,Failure,Diagnosis
