| Title |
Simulation of Depth-averaged Shallow-water Flow with Turbulent Diffusion |
| DOI |
https://doi.org/10.12652/Ksce.2025.45.6.0743 |
| Keywords |
천수 방정식; 수심-평균 흐름 모형; 근사 Riemann 해법; 난류 확산; 수심-평균 k-ε 모형 Shallow water equations; Depth-averaged flow model; Approximate Riemann solver; Turbulent diffusion; Depth-averaged k-ε model |
| Abstract |
To improve the applicability of the flow model for shallow water flow with large streamline curvature, a turbulence model was introduced to reflect eddy viscosity effects. The depth-averaged k-ε turbulence model was combined with the existing depth-averaged 2D shallow-water flow model to account for diffusion by turbulent stresses. To test how well the newly developed turbulence model and the existing flow model worked together, the implemented code was verified and validated by comparison with the laminar analytical solution and turbulent DNS results for Poiseuille flows. The developed model was applied to the abrupt width expansion and the spur dike experiments, where flow is deflected and recirculation flow occurs due to plane geometry, such as a sudden change in channel width. In the simulation of the abrupt width expansion experiment, the turbulent diffusion effect due to eddy viscosity, compared to molecular viscosity, appeared as a reduction of the recirculation zone. When the turbulence model was adopted, the recirculation length scale and the simulated longitudinal flow velocity profiles generally matched the measurements well. In the spur dike experiment where larger discharges were supplied to a narrower channel, it was difficult to properly reproduce the flow with molecular viscosity alone. When turbulent diffusion was considered, the simulated flow velocity profiles were within the vertical range of the measurements even with
the increased discharge, and the RMS errors actually decreased. Thanks to the k-ε turbulence model, it was confirmed that the depth-averaged 2D simulation results for the recirculation flow agreed better with the laboratory experiment results than before. |