摘要
为了有效监测化工厂等场所的危险气体和解决金属氧化物传感器普遍存在交叉敏感性的问题,首先使用不同的MEMS气体传感器组成传感器阵列。然后配制不同的实验气样进行测试,得到实验测试数据,并整理成训练集和测试集样本。最后,采用麻雀搜索算法优化的BP神经网络(SSA-BP)完成气体的定性、定量分析。实验测试结果表明:SSA可以有效提高预测模型的预测精度和稳定性,对乙醇、甲烷、氨气定性识别的正确率达到100%,气体定量预测的最大相对误差不超过5.50%,预测效果得到明显改善。该系统可以满足混合气体的定性和定量分析要求,在危险化学气体监测方面具有良好的应用前景。
In order to effectively monitor dangerous gases in chemical plants and solve the problem of cross-sensitivity of metal oxide sensors,different MEMS gas sensors are first used to form sensor arrays.Then different experimental gas samples are prepared for testing,and the experimental test data are obtained,which are sorted into training set and test set samples.Finally,BP neural network optimized by sparrow search algorithm(SSA-BP)is used to complete qualitative and quantitative analysis of gas.The experimental results show that SSA can effectively improve the prediction accuracy and stability of the prediction model,the accuracy of qualitative identification of ethanol,methane and ammonia reaches 100%,and the maximum relative error of quantitative prediction of gas is less than 5.50%,which significantly improves the prediction effect.The system can meet the requirements of qualitative and quantitative analysis of mixed gas and has a good application prospect in monitoring hazardous chemical gases.
作者
董常庆
施云波
牛昊东
王天
康林贵
李萍
DONG Changqing;SHI Yunbo;NIU Haodong;WANG Tian;KANG Lingui;LI Ping(School of Measurement Control Technology and Communications Engineering,Harbin University of Science and Technology,Harbin Heilongjiang 150080,China;The Higher Educational Key Laboratory for Measuring and Control Technology and Instrumentations of Heilongjiang Province,Harbin University of Science and Technology,Harbin Heilongjiang 150080,China)
出处
《传感技术学报》
CAS
CSCD
北大核心
2022年第8期1093-1101,共9页
Chinese Journal of Sensors and Actuators
基金
国防基础科研计划项目(JCKY2017412C003)
国家自然科学基金项目(62271176)。
