基于神经网络理论的河道水情预报模型

Application of the Neural Network Theory to the Flood Prediction

河道水流运动过程特别是洪水演进过程是一个复杂的非线性动力学过程 ,鉴于神经网络具有很强的处理大规模复杂非线性动力学系统的能力 ,本文将神经网络理论用于河道水情预报的研究 ,以期识别水流运动变化过程与其影响因子之间的复杂非线性关系 ,为河道水情预报提供了一条新的途径。在此基础上建立了螺山站洪水预报的非线性动力学模型 ,通过分析研究得出近年来特别是

Flood evolution exhibits a complicated non linear dynamical process. The neural network possesses the capability of dealing with complex non linear dynamical systems, this paper demonstates how it can be used in flood prediction as a new approach considering the non linear relationship between flood evolution and its factors such as discharge, channel deformation, and so on. Based on it, the neural network approach is applied to the flow prediction of Yangtze River at Luoshan station. The preliminary results suggest that the phenomenon of small discharge with high level in middle reaches of Yangtse River recently, especially in 1998, is related to the downstream aggregation. And the quantitative relations between the water level variation of Luoshan station and the downstream aggregation are obtained.