基于灰色神经网络的作物需水量预测模型研究

Crop water requirement forecasting model based on grey neural network

采用灰色系统理论与神经网络,通过多维度气象因子和参考作物需水量的相关分析,来确立灰色神经网络拓扑结构及网络流程,建立了预测作物需水量的灰色神经网络模型.以海南省儋州市1979～2014年的气候数据为输入,作物需水量为输出数据,运用Matlab工具,仿真表明预测曲线与参考作物需水量曲线拟合度较高.灰色神经网络模型预测结果绝对相对误差均值为5.28％,预测精度高,为节水灌溉提供了一种新的有效方法.

Using the grey system theory and neural network, through the correlation analysis of the multidimensional meteorological factors and crop water requirement, the grey neural network topology structure and the network procedure were established, and the gray neural network model for predicting the crop water requirement was set up. Using the climate data of Danzhou city in 1970-2005 as the input data, the crop water requirements as the output data, the simulation was conducted by the Matlab tool. The fitting degree of the predicted curve and measured curve of crop water requirement is higher, and the mean absolute relative error of the results predicted by the gray neural network model and the actual crop water requirements is 5.27%, which is of high predicting accuracy, and provides a new effective method for the water-saving irrigation.