【摘要】 运用人工神经网络技术的基本原理,以降雨量作为基本影响因子,建立了流域年径流量的神经网络预报模型.在建模过程中,为保证计算快速收敛,重新定义了权重增量的计算公式.从两个流域的应用实例资料验证表明,模型基本合理、可靠,具有较好的适应性和预报精度.由模型计算结果可以看出,将人工神经网络技术应用于流域年径流量的预报研究,是以系统的观点将降雨与径流作为输入和输出联系起来,它可为流域径流的中长期变化预测提供一条崭新而有效的途径.更多还原

【Abstract】 In this paper,by using the principle of artifical neural networks and taking the rainfall as the main affecting factors,a neural network model which can forecast the basin’s annual flow is established.In order to assure the quick astringency of calculation,the weight increment equation is redefined in modeling.The results identified by two basins’s observed data indicate that the model is satisfactory,good adaptability and accuracy.At a viewpoint of system,the artifical neural network technique used for the forecasting of basins’s annual flow can relate the rainfall and flow,considering rainfall as input and flow as output.The model can offer a new and effective method for forecasting the medium and long-term change of basins’s flow.更多还原