摘要
在研究温度对土壤水分入渗的影响时,引入BP网络理论。在5℃、17℃、29℃和40℃4个温度下进行水平一维入渗实验,观测各温度下湿润锋运移距离和累积入渗量随时间的变化;根据5℃、17℃、40℃的湿润锋运移距离和累积入渗量与时间的幂函数关系,以相等的时间间隔收集训练样本,建立土壤水分运动的BP网络模型;用29℃的相应参数为检验样本,对所建模型进行校核,结果良好;该模型可以预测西安粉土5℃～40℃的水分入渗状况,这可以为制定合理的灌溉制度提供重要依据。
The Backpropagation learning algorithm was introduced to investigate the effect of temperature on soil water movement. An experiment of 1D horizontal infiltration at temperature of 5℃, 17℃, 29℃ and 40℃ was carried out. A BP network model was established according to the power function relationships between wetting front distance／cumulative infiltration depth and temperature. The model can be used to forecast soil water movement under a constant temperature. It can be referred as an effective tool for preparing irrigation system.
出处
《水利学报》
EI
CSCD
北大核心
2003年第2期6-8,14,共4页
Journal of Hydraulic Engineering
