基于灰色关联度和BP神经网络的丹江口水库水面蒸发模型研究

Research on Water Surface Evaporation Model of Danjiangkou Reservoir Based on Grey Correlation Degree and BP Neural Network

摘要: 丹江口水库是南水北调中线工程水源地，准确模拟丹江口库区水面蒸发量具有重要的意义。通过收集丹江口水库蒸发站气象及蒸发实测资料，基于灰色关联度分析与BP神经网络相结合的方法，探讨不同气象因子与水面蒸发的关联度，筛选出主要影响因子，并将其作为BP神经网络输入层，构建了基于BP神经网络的丹江口水库水面蒸发模型。结果表明：BP神经网络模型在训练期和验证期，纳什系数分别为0.88、0.81，平均相对误差分别为11.5%、12.6%，所建模型能够准确地模拟丹江口水库的水面蒸发。研究结果可为丹江口水库水资源综合利用和科学调度提供支撑。

Abstract: Danjiangkou Reservoir is the water source of the South-to-North Water Diversion Project, and it is of great significance to simulate the water surface evaporation in Danjiangkou Reservoir area accurately. By collecting meteorological and evaporation measured data from Danjiangkou reservoir evaporation station, based on the method of combining grey correlation analysis and BP neural network, this paper explored the correlation between different meteorological factors and water surface evaporation, screened out the main influencing factors, and constructed water surface evaporation model of Danjiangkou reservoir with BP neural network. The results show that the BP neural network model can accurately simulate the water surface evaporation of Danjiangkou Reservoir. Nash coefficients are 0.88 and 0.81 with average relative errors 11.5% and 12.6% in the training and Validation periods, respectively. The results of the study can provide support for the comprehensive utilization of water resources and scientific scheduling of Danjiangkou Reservoir.