中文摘要：

介绍了采用热催化传感器在不同的工作温度下,通过BP人工神经网络对可燃混合气体进行分析的新方法.研究和建立了神经网络的结构模型和学习算法,并以3种混合气体为对象进行仿真训练,训练结果可以较好地实现对可燃混合气体的分析.

英文摘要：

A new method of analysis to mixed inflammable gas with a heat catalytic sensor worked in different conditions through BP artificial neural network is introduced. Structural model and algorithm are also studied and established. Finally three mixed gases are used as object to be simulated. The result shows that the method can identify and analyze the ingredients of mixed gas accurately.