基于BP神经网络的灌区耗水量模拟预测模型

青铜峡灌区耗水量是一个复杂的非线性系统,受多种因素的影响,而农业耗水量是耗水大户.将影响灌区农业耗水量的主要因素:引水量、排水量、降雨量、蒸发量和地下水位埋深作为影响因子,建立青铜峡灌区灌溉耗水量的BP神经网络预测模型.结果表明,应用此网络,对2001-2003年灌区灌溉耗水量进行了预测,误差满足要求,精度较高,说明网络函数的选取和结构设计较为合理,该模型可用于灌区农业耗水量的预测.

The simulating and forecasting of water consumption model base on BP neural network in irrigation area

Qingtongxia Irrigation Area water consumption is a non-linear system,which suffers many kinds of influencing factor restriction,but the agricultural water consumption is the most important component part.In view of the influencing factor of agriculture water consumption in the irrigation area,the main impact factors are the irrigation water volume,the drainage water volume,the rainfall volume,evapotranspiration and ground water table,establishes the BP neural network forecast model of agriculture irrigation water consumption in Qingtongxia eastern irrigation area.The result indicates that selection of network function and the structural design are more reasonable,the error satisfies the needs;the precision is high,which may apply in the forecast of water consumption in other Area.Using this network,forecast the agriculture irrigation water consumption in Qingtongxia irrigation area in 2001-2003.The fact proves that this model suits the prediction of agriculture water consumption in the irrigation area.