摘要
针对本实验室设计的红外甲烷传感器,在实验过程中易受到传感器内部干扰信号影响的问题,使得传感器产生较大的测量误差,为此提出了一种基于TensorFlow架构的Adam-BPNN进行传感器系统误差修正的方法。实验与数据处理结果表明:通过该算法的误差修正,能有效地减小传感器内部干扰信号对测量结果的影响,相对误差减小到0.0226%,提高了红外甲烷传感器的稳定性与精度,且重复性好。
Aiming at the problem,which the infrared methane sensor is susceptible to the internal interference signal,may cause a considerable measurement error.In this paper,a method of systematic error correction using Adam-BPNN based on the TensorFlow architecture is purposed.The results show that the influence of the internal signal on the measurement result can be effectively reduced by the correction of this algorithm.The relative error is reduced to 0.0226%,which improves the stability and accuracy of the sensor,and has excellent repeatability.
作者
陈红岩
盛伟铭
刘嘉豪
黄翰
朱俊江
赵永佳
CHEN Hongyan;SHENG Weiming;LIU Jiahao;Huang Han;Zhu Junjiang;ZHAO Yongjia(China Jiliang University College of Modern Science and Technology,Hangzhou 310018,China;College of Mechanical&Electrical Engineering,China Jiliang University,Hangzhou 310018,China)
出处
《传感技术学报》
CAS
CSCD
北大核心
2020年第4期529-536,共8页
Chinese Journal of Sensors and Actuators
基金
浙江省重点研发计划项目(2019C03114)。
