摘要
提出了径流长期分级预报一种新的模式识别方法。该方法不直接描述径流形成的物理机制，而是通过Ｋｏｈｏｎｅｎ自组织神经网络对历史样本（径流级别及其影响因子集）的学习，识别出蕴含在样本中径流级别与其影响因子之间的规律性。辽宁省大伙房水库基于气象因素的汛期径流预报实践，表明了该方法的可行性和有效性。
In this paper,we investigate the long term runoff stage forecast by using Kohonen network.The new methodology is tailored to be applied to the runoff forecasting of Dahuofang reservoir.The results indicate the method is feasible and effective and also show the potential of the new algorithm as an alternative to the traditional methods of hydrological forecasting.
出处
《水电站设计》
1997年第2期24-29,共6页
Design of Hydroelectric Power Station
基金
国家自然科学基金
青年基金
