基于RBF算法的探空湿度太阳辐射误差预测  

RBF algorithm based prediction of solar radiation error detected with radiosonde humidity method

作　　者：冒晓莉[1] 张鹏[1] 张加宏[1] 赵雪伟 MAO Xiaoli;ZHANG Peng;ZHANG Jiahong;ZHAO Xuewei(School of Electronic&Information Engineering,Nanjing University of Information Science&Technology,Nanjing 210044,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

出　　处：《现代电子技术》2020年第19期146-151,共6页Modern Electronics Technique

基　　金：国家自然科学基金资助项目:GTS1探空湿度测量误差修正方法的关键问题研究(41605120)。

摘　　要：针对传统BP神经网络算法预测的探空湿度太阳辐射温度误差偏大的问题,基于南京大桥的GTS1-2湿度传感器及其防雨帽模型,采用计算流体动力学(CFD)软件,通过PRO/E建模、ICEM划分网格及FLUENT仿真,以高空实际探测中典型气压、太阳高度角和太阳辐射量为变量仿真出2530组温度误差的数据样本。数据样本通过BP,PSO-BP,GA-BP,RBF神经网络算法进行优化对比,最终采用RBF神经网络算法构建预测模型,可预测出不同环境下探空的湿度太阳辐射温度误差,且预测出的温度误差最小。In view of the large error of the radiosonde humidity solar radiation temperature predicted by the traditional BP neural network algorithm,the CFD(computational fluid dynamics)software,including PRO/E modeling,ICEM meshing and FLUENT simulation is adopted to simulate 2530 groups of data samples of temperature error on the basis of the GTS1-2 humidity sensor and its rainproof cap model of Nanjing Bridge.In the simulation,the typical air pressure,solar altitude and amount of solar radiation in high altitude actual detection are taken as variables.The data samples are optimized and compared with BP(back propagation),PSO-BP(particle swarm optimization-back propagation),GA-BP(genetic algorithm-back propagation)and RBF(radial basis function)neural network algorithms.Finally,the RBF neural network algorithm is used to build the prediction model,which can predict the radiosonde humidity solar radiation temperature error in different environments,and the error of the predicted temperature is the minimum.

关 键 词：气象探测 GTS1-2湿度传感器 计算流体动力学 太阳辐射偏干误差 粒子群优化神经网络算法 遗传神经网络算法 径向基函数神经网络算法 

分 类 号：TN911.1-34[电子电信—通信与信息系统] TP183[电子电信—信息与通信工程]