基于MATLAB7的神经网络在灌溉用水量预测中的应用研究

Neural Network's Application on the Forecast of Irrigation Water Use Based on MATLAB7

灌溉用水量的预测对于灌区管理工作具有重要的指导意义,使用神经网络方法预测灌溉用水量。介绍了BP网络的算法步骤,并且以铁甲灌区为例,使用软件MATLAB7对所设计的网络进行学习和训练,隐含层单元数的选取采用实验法,最终以隐单元数为13的网络预测性能最好,误差也达到精度要求。所建模型可以预测铁甲灌区的灌溉用水量。

The forecast of irrigation water use has important guiding role to the management of irrigation. This pooper used neural network method to forecast irrigation water requirement. It introduced BP network algorithm, taking Tiejia irrigation area as an example, used the software MATLAB 7 to study and train the designed network. The selection of concealed level single number used the cut-and-try method. Finally the forecasting performance of single number of 13 was the best and the error meets the precision requirements. The model can be used to predict irrigation water use of Tiejia irrigation area.