模糊神经网络在发动机故障诊断中的应用研究

An application study of the fuzzy neural network used in an engine fault diagnosis system

对模糊逻辑和神经网络故障诊断方法做了研究,根据这两种方法各自的优缺点,采用串联方法将两者相结合,用模糊信息处理方法对输入信号进行预处理,然后利用神经网络的逼近能力来实现对故障的诊断。将该方法构建的推理系统应用于汽油发动机偶发性疑难故障诊断,仿真结果表明该方法可以给出较高精度的诊断结果。

This paper deals with a study of both fuzzy theory and neural network technology in a fault diagnosis. The strong points of the two methods are combined with each other. The fuzzy method is used to process the information and the fault diagnosis is implemented by the approaching ability of the neural network. An inference system is set up by using this method to solve complicated faults occurring in a gasoline engine. The simulation result has proved that the system can achieve an accurate fault diagnosis.