【摘要】 由于水沙作用机制和演进规律的复杂性 ,以及河道形态变化等因素 ,多泥沙洪水预报一直是洪水预报的难点 ,对高含沙洪水快速、准确的预报是多年来国内外专家十分关注的课题。作者采用具有高度非线性识别能力的人工神经网络与遗传算法相结合的方法 ,探讨了建立智能预报模型的基本方法 ,进一步对如何提高预报精度的问题进行了研究 ,并结合黄河洪水预报实例检验了神经网络模型的可行性。检验结果表明 ,该方法能够较好地识别多泥沙洪水的演进规律 ,对水位、流量和含沙量都能进行合理预报更多还原

【Abstract】 For many years, a lot of domestic and foreign experts have been concentrating on such an issue that how to forecast the silt\|laden flood quickly and accurately because silt\|laden flood forecasting is a very difficult problem which is caused by the complicated mechanism of water and sediment movement and the channel bed variation.This paper discusses the fundamental methods how to establish an intelligence forecasting model, which is based on the methods combined the artificial neural networks, which has the high non\|linear identifying ability, with the genetic algorithms. In addition,how to improve forecasting precision is investigated and the feasibility of applying this neural networks model to the Yellow River flood forecasting is tested.The testing results prove that this method can identify the law of evolution of silt\|laden flood better and it can forecast the water level, discharge and Sediment concentration reasonably.更多还原