BP神经网络在流溪河水库径流量预测中的应用

Application of BP Neural Network Model in Forecasting Runoff of Liuxihe Reservoir

流溪河水库是广州市的主要饮用水源地,所属地区降雨量年际变化大,不利于水资源的优化配置。对流溪河水库径流量进行预测研究,可为其水资源的优化配置提供科学依据。目前人工神经网络（ANN）技术在水文序列模拟预测中有较多的应用,本文根据流溪河水库1959-2000年水文数据,利用BP神经网络对径流量进行预测,从模型检验结果看,所建模型有较好的拟合效果和预测精度,说明神经网络在预测径流量方面有良好的实用性。

Liuxihe Reservoir is the major drinking water source for Guangzhou,but the interannual variation of precipitation changed obviously,which has deterimental effects on the management of water resources.Forecasting its runoff can provide a scientific decision for future water allocation.Now,artificial neural network（ANN）model has been applied to forecast and simulate the regular pattern of the hydrologic time series.In this paper,the ANN model uses the hydrological data from 1959 to 2000 to forecast runoff.The forecasting results show that the model has good adaptability and precision,and the ANN model has good practicability in forecasting runoff.