【作者】 周平；

【导师】 芮孝芳；

【作者基本信息】 河海大学 ， 水文学及水资源， 2007， 硕士

【摘要】 扩散波洪水演算法是一种既具有足够精度又便于求解的方法。本文对扩散波洪水演算进行了较为系统的研究，着重探讨了其数值差分解。基于运动波数值扩散在一定条件下可以模拟扩散波的物理扩散，探讨了马斯京根—康吉洪水演算法和一种具有预见期的差分解法；分析了方法的相容性、稳定性和收敛性，讨论了采用不同方式处理边界的相容性、稳定性和收敛性条件；介绍了运动波差分方程模拟扩散波方程近似解的精度准则，给出了二阶、三阶和四阶精度解的最优条件。利用直接差分格式的多样性，给出了扩散波方程的两种显式直接差分解、一种隐式直接差分解和三种混合直接差分解，讨论了以上6种差分方程及其边界处的相容性、稳定性和收敛性条件。对汉江襄阳至皇庄河段和西江梧州至高要河段，应用扩散波解析解法、马斯京根—康吉洪水演算法、具有预见期的差分解法和6种直接差分法对其进行洪水演算；采用遗传算法根据一定的准则对训练样本的各场次洪水的模型参数进行率定，进而利用信息扩散技术由训练样本的模型参数获得校核样本的模型参数；并给出通过不同方法获得的模型参数的校核样本的预报结果。最后，比较了不同方法的扩散波洪水演算结果，结果表明，不同方法对两条河段的演算精度均较高，通过优化和信息扩散方法得到的样本模型参数好于直接采用物理方法获得模型参数的计算结果。更多还原

【Abstract】 It is known that the diffusion-wave flood routing method is a good one with accuracy and simplicity. In this thesis, this method is systematically studied, especially the numerical difference solution to the equations.Based on principle that numerical dispersion of the linear kinematic-wave can simulate physical dispersion of the diffusion wave on certain condition, Muskingum-Cunge method and a flood routing method with forecasting period are discussed; the compatibility, stability and convergence of the two methods are analyzed, and the conditions to handle the compatibility, stability and convergence at the boundary in different methods; accurate principles of the approximation of that simulation are introduced, and optimal conditions necessary to obtain second-, third-and fourth-order accurate solutions are also given. By making use of the diversity of the finite difference scheme, two explicit direct difference, one implicit direct difference and three mixed direct difference schemes are given, and the above six kinds of difference equations together with their compatibility, stability and convergence conditions at the boundary are discussed.In the case of Xiangyang-Huangzhuang channel (part of Hanjiang river of Changjiang) and Wuzhou-Gaoyao channel (part of Xijiang river of Zhujiang), methods of the diffusion wave analysis and Muskingum-Cunge, flood routing method with forecasting period, and six finite difference solutions are applied to the calculation of the diffusion wave; according to certain rules, real coding based accelerating genetic algorithm is used to optimize the model parameters of studied samples, from which the model parameters of the verified samples are obtained by the way of information diffusive technique; in addition, the forecasting outcome the model parameters of the verified samples through different methods are given.After comparing the results of different methods, the outcome shows that all methods have satisfactory precision, and model parameters of samples gained by optimization and information diffusive technique are better than those by physical means.更多还原