A comparative experimental and multiphysics computational fluid dynamics study of coupled surface–subsurface flow in bed forms

Abstract

The use of multiphysics computational fluid dynamics (CFD) approaches to simulate surface–subsurface flow processes is evaluated by comparison with flume experiments on current‐exposed permeable bed forms. The unique experimental data include measurements of the time‐averaged surface water flow velocities, the pressure distribution at the sediment–water interface, and pore water flow paths. The modeling approach first simulates the time‐averaged turbulent flow in the channel with CFD and then uses the predicted pressure distribution at the sediment–water interface to drive a flow and transport model for the sediment. The CFD‐modeled velocity and pressure distribution and transient particle tracks within the sediment agree reasonably well with observations. Differences that exist between observations and simulations mainly concern the eddies in the wake zone downstream of the ripple crests that are slightly shorter than those predicted by the model. This deviation propagates from the surface to the subsurface domain, appearing in the pressure distribution along the bed and, consequently, the subsurface flow patterns. The good representation of general patterns and rates makes multiphysics CFD modeling a powerful and sufficiently accurate tool that can replace measurements for many studies of surface–subsurface processes involving current‐exposed immobile bed forms. The approach can be used for predicting transport processes where they cannot easily be observed, such as in large rivers and coastal systems where boundary conditions such as mean currents and bed forms can be mapped.