科研之友
微信
新浪微博
Facebook
分享链接
摘要
The shallow water flow solver within Gerris Flow Solver implements second-order accurate Gudunov type numerical schemes, with preserving the balance of source and flux terms on quadtree cut cell grids. The solver provides flexible grid generation, with using adaptive quadtree grids and the cut cell method, which can improve computational efficiency in 2D simulations. The solver, however, cannot implicitly discretize nonlinear friction without linearization since the current version of Gerris supports the implicit discretization only for linear friction. We compare implicit treatment for Manning's friction to the linearization. Simulation results using a benchmark test show the difference between two methods is only significant when the roughness coefficient is unrealistically high. Two real flood events, Malpasset dam break in France and Baeksan levee failure in Korea, are simulated using the quadtree grid based shallow water model, with adaptively refining meshes near water fronts and river boundaries. Comparison of simulation results with observations and measurements demonstrates that the grid adaptation can save approximately 85%-95% of the computational cost while preserving the accuracy. The research result shows that the linearized friction using the current version of Gerris can be applied for flood simulation and multiple inflow boundaries can be treated as source terms.
- 出版日期2015-10-14
