Explicit Accelerated-Convergence Algorithms for Steady Shallow Water Flows

Computations of steady shallow water flows are often perceived to be more difficult than unsteady ones due to low-rate convergence to steady states. This paper explores the possibility of utilizing five methods from CFD to accelerate the convergence, with some elaborate modifications peculiar to water flows. Based on numerical experiments, local time-stepping, redidual smoothing, and multigrid methods are recommended, in conjunction with the use of a critical value of CFL number determined through trial computations. By choosing an appropriate approach for a specific problem, a 30%-50% or even larger reduction in computational effort can be achieved.