基于Elman网络的黄河源区枯期径流预报

The Yellow River Dry Season Prediction Research Based on Elman Neural Network

采用反馈Elman网络，对黄河唐乃亥水文站及上游玛曲站近45年（1959—2003年）年降雨及径流流量进行分析，建立了基于反馈神经网络的黄河源区枯季径流预报模型．利用Matlab7神经网络工具箱对黄河源区唐乃亥水文站枯季径流量进行了预报．结果表明所建立的ANN（7，7，15，7）模型预报结果精度高，容错能力强，是枯季径流预报的有效手段．

According to the annual and run- off precipitation in the successive 45 years（ 1959 - 2003）at Tangnaihai and Maqu Hydrology Station in the upper reach of the Yellow River, a Elman neural network model of dry season run - off forecasts is established based on the principle of feedback artificial neural network. Then the dry season prediction of Tangnaihai Hydrology Station of Yellow River is given by Neural Network Toolbox of Matlab7. The results of the training and test show that the ANN（7,7,15,7）model based on Elman is more precise,so it is feasible to be applied for the forecast of dry season forecast.