基于径向基函数网络的钱塘江涌潮模型及其应用

Radial basis function network based tidal bore model and its application to Qiantang estuary

利用径向基函数(RBF)网络的最佳函数逼近性能,建立了盐官站涌潮潮头高度的预测模型.钱塘江涌潮与其影响因素之间存在复杂的非线性关系,模型的预测精度主要取决于影响因子的识别与训练样本的选择.实测资料检验表明该模型具有良好的泛化能力.应用该模型分析了在不同水动力与河床边界条件下,大规模治江围涂工程对盐官潮头高度的影响.结果表明,在不同河势条件下治江围涂工程引起的涌潮变化存在差异,对于顺直江道,盐官大潮潮头高度降低了0.17 m,而对于弯曲江道,则抬高了0.39 m.

A model for predicting the bore height at Yanguan was established by using the function approximation performance of radial basis function （RBF） network. The model identifies the complex nonlinear relationships between the tidal bore and its influencing factors, and model prediction accuracy depends on identifying the major influencing factors and improving the representative learning samples. Comparison of the simulation results with the observation data showed that the model had good generalization ability. Then the model was used in a case to study the effect of large-scale reclamation and regulation works on the tidal bore in the Qiantang estuary. The influence extent to the bore height at Yanguan was quantitatively analyzed under different hydrodynamic and river-bed condition. The results showed that the bore height decreased by 0. 17 m after reclamation under straight alignment of river course, and increased by 0.39 m under curved one,