基于BP神经网络的农田灌溉量预测  

Forecast of Amount of Farmland Irrigation Based on BP Neutral Network

作　　者：沈岳[1] 曾文辉[2] 欧明文 匡迎春[2] 

机构地区：[1]湖南农业大学信息科学技术学院,长沙410128 [2]湖南农业大学工学院,长沙410128

出　　处：《农机化研究》2015年第4期36-39,共4页Journal of Agricultural Mechanization Research

基　　金："十二五"国家科技支撑计划项目(2012BAD35B05)

摘　　要：农田灌溉量预测对水资源合理规划和优化配置具有重要意义。为此,以农田水分平衡公式为基础,导出灌溉量I与前各时期的降雨量pt-1、pt-2、pt-3灌溉面积S之间存在非线性关系,由此建立了农田灌溉量的BP神经网络预测模型。检验结果表明,灌溉量BP神经网络预测法精度约为94.5%。因而,以灌溉面积、前各时期降雨量为输入变量,结合BP神经网络技术来预测农田灌溉量,具有很好的应用前景。Prediction of farmland irrigation is of great significance to rational planning and optimal allocation of water resources .The article based on the farmland water balance equation ,deduced that nonlinear relationship between irrigation amount I and pt-1 、pt2 、pt-3 that rainfall irrigation amounts of each period、irrigation area S;which established a BP network prediction model irrigation amount nerve .Test results show irrigation amount BP neural network prediction accuracy is about 94 .5%.Thus we can see that in the irrigated area , rainfall before each period as input variables, combined with BP neural network techniques to predict the amount of irrigation, has good prospects.

关 键 词：农田灌溉量 降雨量 预测模型 BP神经网络 

分 类 号：S274.4[农业科学—农业水土工程]