基于脉冲供电的催化燃烧式气体传感系统  

Catalytic combustion gas sensing system based on pulse power

作　　者：刘文鹏 陈向东[1] 吴宇尘 LIU Wen-peng;CHEN Xiang-dong;WU Yu-chen(School of Information Science and Technology,Southwest Jiaotong University,Chengdu 611756,China)

机构地区：[1]西南交通大学信息科学与技术学院,成都611756

出　　处：《信息技术》2020年第8期1-6,11,共7页Information Technology

基　　金：国家自然科学基金重点资助项目(61731016)。

摘　　要：催化燃烧甲烷气体传感器对于高浓度甲烷检测的灵敏度往往较低。文中设计了一种脉冲式供电和多传感器信息融合技术相结合的甲烷检测系统,采用多个传感器获取环境中与催化燃烧反应有关的湿度、温度、甲烷、氧气和二氧化碳等参数,再将获取的传感器数据通过局部异常因子(Local Outlier Factor,LOF)算法进行预处理。在此基础上,通过随机森林(Random Forest,RF)算法提取重要特征,再去除共线性特征,然后通过极端随机树(Extremely Random Forest,ERF)算法对多个传感器采集的数据进行训练得出预测模型,以便实现较高浓度的甲烷预测。Catalytic combustion methane gas sensors are often less sensitive to high-concentration methane detection.In this paper,a methane detection system combining pulsed power supply and multi-sensor information fusion technology is designed.Multiple sensors are used to obtain the humidity,temperature,methane,oxygen,and carbon dioxide parameters related to the catalytic combustion reaction in the environment.The sensor data is pre-processed by a local outlier factor(LOF)algorithm.On this basis,important features are extracted by random forest algorithm(RF),collinearity features are removed,and then data collected by multiple sensors are trained by extreme random tree(ERF)algorithm.A prediction model is developed to enable prediction of higher concentrations of methane.

关 键 词：脉冲供电 甲烷检测 催化燃烧 STM32F051 随机森林 

分 类 号：TN402[电子电信—微电子学与固体电子学]