基于BP神经网络的温度传感器辐射误差修正  

Radiation error correction of temperature sensor based on BP neural network

作　　者：吴翼凡 刘清惓[1,2,3] 杨杰 戴伟[4,5] WU Yifan;LIU Qingquan;YANG Jie;DAI Wei(School of Electronics and Information Engineering,Nanjing University of Information Science&Technology,Nanjing 210044,China;Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology,Nanjing 210044,China;Jiangsu Key Laboratory of Meteorological Observation and Information Processing,Nanjing 210044,China;Key Open Laboratory for Aerosol.Cloud.Precipitation of China Meteorological Administration,Nanjing 210044,China;Key Laboratory of MEMS of the Ministry of Education,Southeast University,Nanjing 210096,China)

机构地区：[1]南京信息工程大学电子与信息工程学院,江苏南京210044 [2]江苏省大气环境与装备技术协同创新中心,江苏南京210044 [3]江苏省气象探测与信息处理重点实验室,江苏南京210044 [4]中国气象局溶胶与云降水重点开放实验室,江苏南京210044 [5]东南大学MEMS教育部重点实验室,江苏南京210096

出　　处：《现代电子技术》2018年第24期43-47,共5页Modern Electronics Technique

基　　金：国家公益性行业(气象)科研专项(GYHY200906037,GYHY201306079);国家自然科学基金(41275042);江苏高校优势学科Ⅱ期建设工程资助项目(PAPD-Ⅱ)。

摘　　要：针对探空仪搭载的温度传感器受太阳辐射加热的影响而产生测量温度误差的问题,提出一种将BP神经网络应用于温度传感器辐射误差的预测方法。将海拔高度、探空仪上升速度、太阳辐射强度作为BP神经的输入,利用计算流体力学分析该传感器在不同环境条件下的辐射误差作为BP神经网络训练的预测模型,分析预测结果与样本数据,两者的绝对误差区间为[-0.003,0.005],测量精度达到0.1 K量级的要求。用C#和Java将BP神经网络得到的辐射误差修正方程分别进行封装,并完成Windows上位机和安卓端移动应用开发。In allusion to the temperature measurement errors caused by heating of solar radiation acting on the temperature sensor carried on a radiosonde,a method of applying BP neural network to radiation error prediction of the temperature sensor is proposed.The altitude,rising speed of the radiosonde and solar radiation intensity are taken as the inputs of the BP neural work.The computational fluid mechanics is used to analyze the radiation errors of the sensor in different environmental conditions,which are taken as the prediction model for training of the BP neural network.The prediction results and sample data are ana-lyzed.The results show that the absolute error margin is[-0.003,0.005],and the measurement accuracy of the temperature sen-sor can reach the order of 0.1 K.The C#and Java languages are respectively used to encapsulate the radiation error correction equations obtained by the BP neural network,and accomplish the development and mobile applications of the Windows upper computer and Android terminal.

关 键 词：温度传感器 探空仪 流体力学 BP神经网络 辐射误差 误差修正 

分 类 号：TN915-34[电子电信—通信与信息系统] TP274[电子电信—信息与通信工程]