基于BP网络的场次洪水冲淤量计算

Calculation of volume of scour/deposition during a separate flood based on BP neural network

本文针对黄河下游1950-1985年298场历史洪水实测资料进行了统计分析,分析出了影响场次洪水冲淤量的主要因素,利用BP神经网络建立了各影响因素与下游河道冲淤量的关系模型,并对模型进行了验证,结果表明:计算值与实测值吻合较好,计算精度较高,可以用来对场次洪水冲淤量进行预报,计算方法较传统方法简便。

In this paper the data from the 298 flood events during the period of 1950-1985 are analyzed using statistical method.The related factors of the volume of scour /deposition during a separate flood are obtained and the model is built by BP neural network.The volume of scour /deposition is calculated and predicted by this model.The results of BP neural network are in good agreement with the observed values and the average errors of calculation and prediction are quite small.So this model can be used to estimate the volume of scour / deposition durring a separate flood and the calculation is more convenient than traditional methods.