基于传感器阵列的气体检测与分析系统设计  

Design of gas detection and analysis system based on sensor array

作　　者：喻璐 谭志文 邹望辉 YU Lu;TAN Zhiwen;ZOU Wanghui(School of Physics and Electronics,Changsha University of Science and Technology,Changsha 410114,China)

机构地区：[1]长沙理工大学物理与电子科学学院,湖南长沙410114

出　　处：《电子设计工程》2022年第10期129-133,138,共6页Electronic Design Engineering

基　　金：湖南省自然科学基金(2020JJ46)。

摘　　要：工业生产环境中存在各种严重危害人体健康的气体,其中比较常见的有乙醇和丙酮。该设计搭建了一个基于传感器阵列的电子鼻系统,用于乙醇和丙酮的快速检测,可以实现对乙醇、丙酮混合气体的分类。为研究不同的阵列组合,设计了阵列PCB板,分别采集了单一传感器TGS2620与传感器阵列在0~200 ppm目标气体浓度范围内的阻值数据,按照分段拟合的方式进行特征提取,并结合带高斯核的支持向量机(SVM)模型来实现分类。结果表明,该模型能有效识别出乙醇和丙酮,单一传感器条件下的分类正确率为77.1%,传感器阵列条件下的分类正确率为92.5%,传感器阵列条件下识别的精度显著增高,具有明显优势,能够满足工业生产检测需求。In industrial production environments,there exists a variety of gases,such as ethanol and acetone,which could cause serious problem to the human health.This paper provides a design of electronic nose system based on sensor array,which can be used for rapid detection and classification of ethanol and acetone.In order to study various combination of sensor array,a PCB board is designed.The resistance data sets for single sensor TGS2620 and sensor array configuration are both collected in a gas concentration range of 0~200 ppm.Segment fitting method is used for feature extraction,and Support Vector Machine(SVM)model with Gaussian kernel is applied to realize mixed gas classification.The results show that the classification accuracy of single sensor is 77.1%,and the classification accuracy of sensor array is 92.5%.The SVM model with Gaussian kernel can effectively identify ethanol and acetone,the accuracy of sensor array recognition is significantly improved,and it can meet the needs of industrial production testing.

关 键 词：电子鼻系统 传感器阵列 支持向量机 特征提取 

分 类 号：TN98[电子电信—信息与通信工程]