摘要
依据乌鲁木齐市气象站多年来的气象数据,利用改进BP神经网络建立太阳辐射日总量与日照时数的关系模型,并与一些常用函数的曲线拟合结果进行分析比较,以建立有效的基于日照时数的太阳辐射估算模型.利用改进的BP神经网络模型估算日辐射总量,其绝对误差和相对误差分别为0.499 MJ·m-2和3.90%,估算结果与实测值吻合良好.
The relation models between the daily solar radiations and the daily sunshine durations were established by using improved BP neural network based on the meteorological data of many years from the Urumqi weather station. Some other common functions were also applied to take curve fitting between the daily solar radiations and the daily sunshine durations and were compared with the BP neural network model. The daily solar radiations were estimated by the improved BP neural network model and the absolute error and relative error of the estimated value compared with the measured value was 0. 499 MJ . m^-2and 3.90% respectively, which means good agreement was reached between the model and the actual situation.
出处
《河南农业大学学报》
CAS
CSCD
北大核心
2013年第6期732-736,共5页
Journal of Henan Agricultural University
基金
河南省科技攻关项目(122102210128)
关键词
太阳辐射
日照时数
BP神经网络
solar radiation
sunshine duration
BP neural network