该文介绍了一种基于人工神经网络进行气体传感器故障检测的新方法,文中利用单个气体传感器的输出信息为气体传感器建立了动态非线性神经网络气体传感器输出模型,并利 用该模型进行在线故障检测。实际使用证明该模型具有良好的收敛性和稳定性,完全能满足对气体传感器故障在线检测的需要。 [关键词] 气体传感器;故障诊断;神经网络;非线性 Abstract :This paper describes a new method for the ANN2based fault detection of a gas sensor. The dy2 namic nonlinear model2based neural network is established only by the output information of a single gas sen2 sor . The on2line fault diagnosis is made with the help of this model . Since the model is provided with good convergency and stability , it can meet the needs of the on2line detection of a gas sensor. Key words :Gas sensor ; Fault diagnosis ; Neural network ; nonlinear