摘要
催化传感器对不同可燃气体或在不同的工作温度下有不同的输出灵敏度,根据这一特点控制催化传感元件的工作温度,检测可燃混合气体在不同工作温度下的输出信号,由催化传感元件的静态热平衡关系,得出信号与混合气体中各种可燃气体浓度相关的方程组,由于输出信号与气体浓度的关系是非线性的,故采用人工神经网络进行训练,建立了分析多种可燃气体的数学模型.通过对甲烷,丁烷和乙炔3种气体混合的5组样品进行实验,分析结果的绝对误差均小于0.5%,证明所研究的方法可以较好地实现单一催化传感器对多种可燃混合气体的分析.
A catalytic sensor has different sensitivity for different gases or under different temperatures. To control a catalytic sensor worked under different temperatures, the different output signal detecting the inflammable gases can be obtained. By static state heat balance connection of catalytic sensor, a group of equations interrelated output singal and consistence of diversified the inflammable gases in the mixed gases are proposed. The radial basis function （RBF） neural network is used to train them because the of non-linear relation between output and gas consistence. A mathematic model of analyzing the diversiform inflammable gases is established. By the experimental analysis for mixed gases such as firedamp, butane and acetylene. The results show that it is reliable to analyze mixed inflammable gases by a catalytic sensor the absolute error of analyzing result is less than 0.5 %.
出处
《中国矿业大学学报》
EI
CAS
CSCD
北大核心
2006年第1期35-38,共4页
Journal of China University of Mining & Technology
基金
国家自然科学基金资助项目(50374067)
