| 基于BP神经网络的武汉市分层水温预报模型探究 |
| |
| 引用本文： | 陆鹏程,孟翠丽,赵辛慈.基于BP神经网络的武汉市分层水温预报模型探究[J].现代农业科技,2022(6). |
| |
| 作者姓名： | 陆鹏程 孟翠丽 赵辛慈 |
| |
| 作者单位： | 武汉农业气象试验站,武汉农业气象试验站,武汉市气象局 |
| |
| 摘 要： | 为探讨武汉市分层水温的变化特征及预报方法,本文以武汉市金银湖为例,利用2019-2020年地面观测和分层水温资料,分析气温、气压、蒸发、地温、日照及辐射等气象要素对垂向水温变化的影响,建立武汉地区基于拟牛顿法反向传播(back propagation,BP)神经网络的20cm、40cm、60cm及80cm等层次水温预报模型。仿真结果表明：拟牛顿法BP神经网络的水温预报模型能够表达水温和气象要素的非线性关系,平均预报准确率超过90%,具有较高的预报精度。
|
| 关 键 词： | 水温预报 预报模型 神经网络 拟牛顿法 |
| 收稿时间： | 2021/6/11 0:00:00 |
| 修稿时间： | 2021/6/11 0:00:00 |
Exploration of layered water temperature prediction Modelbased on back propagation Neural network in Wuhan city |
| |
| Abstract: | In this paper, author take Wuhan Jinyin Lake as an example, to study the variation characteristics of layered water temperature in wuhan city and forecast method. The influence of meteorological factors, such as air temperature, air pressure, evaporation, ground temperature, sunlight and radiation, on the change of vertical water temperature was analyzed by ground observation and stratified water temperature data in 2019-2020, to establish the 20cm, 40cm, 60cm and 80cm hierarchical water temperature prediction models based on quasi-Newtonian back propagation (BP) neural network in Wuhan area. The simulation results show that the water temperature prediction model based on BP neural network can express the nonlinear relationship between water temperature and meteorological elements, and the average prediction accuracy is over 90%. This model has high prediction accuracy. |
| |
| Keywords: | Water temperature Prediction model Neural network Quasi-newton method |
|
| 点击此处可从《现代农业科技》浏览原始摘要信息 |
| 点击此处可从《现代农业科技》下载免费的PDF全文 |
|
