人工神经网络在气体分析系统中的应用

为了研究人工神经网络在气体分析系统中的应用,选择了识别CH4和CO的传感器阵列及用于网络模型训练的样本。实验分析确定了信号预处理方法,选取合适的网络模型并创新性的对用于气体种类识别的BP算法进行改进,并研究网络结构和参数的变化对气体分析系统性能的影响,实验结果证明气体传感器阵列与BP算法相结合进行气体成分分析具有一定的可行性,若能进一步调整及完善气体传感器阵列的硬件测量电路,则可开发出用于现实复杂环境的多气体分析系统。

Application of Artificial Neural Network in Gas Analysis System

In order to study the application of artificial neural network in gas analysis system,the article chooses gas sensor array and network model training samples to identify CH4 and CO,signal preprocessing algorithms is determined in the experiment.The experiment also chooses appropriate network model and creative improves BP algorithm for gas identification.The article researches the influence on gas analysis system of structure and parameters of feed forward neural network,result indicates that the method based on the combine gas sensor array with BP nervous network to analyze gas component is feasible.If further adjusting and improving hardware measurement circuit of gas sensor array,the multi-gas analysis system,which is applied in reality complex environment can be developed.