Large eddy simulation of two-way coupling sediment transport

Abstract

In the present paper numerical simulations are used to investigate suspended sediment transport and its effect on the dynamics of the turbulent boundary layer. We use an Euler–Euler methodology based on single-phase approach. Large eddy simulation is employed to resolve the large scales of motion, whereas the contribution of the small scales is parametrized by the use of a dynamic Smagorinsky model. In order to account for sediment-induced buoyancy on momentum, a buoyancy term is considered in the three-dimensional Navier–Stokes equations through the use of the Boussinesq approximation. We consider four sediment sizes and the simulations are performed for both one-way and two-way coupling approaches to gain a better description of sediment–turbulence interaction. The level of stratification for each particle size is qualified by the bulk Richardson number which increases by decreasing the grain size. The analysis reveals that the reduction of sediment size produces a larger suspension and sediment concentration in the flow field, due to the concurrence of increased available concentration at the wall and reduced deposition velocity. Comparison of concentration profiles between one-way and two-way coupling clearly shows the remarkable effect of stratification on the velocity and concentration mean profiles. This is particularly true for small sediments which are more likely suspended in the fluid column. In agreement with experimental literature results, our study shows that suspended sediment concentration reduces the von Kàrmàn constant of the velocity profile. The analysis of second order statistics and energy power spectra show turbulent suppression due to stratification effects, in agreement with previous studies. The gradient Richardson number distribution along the channel height demonstrates the increased level of stratification along the fluid column by decreasing the grain size. Momentum and concentration diffusivity are also discussed. The non-dimensional concentration diffusivity compares very well with the Coleman’s experimental data in the range of parameters shared by the two studies. Overall, the results of our study confirm that a single-phase mathematical model is a good candidate to simulate suspended sediment transport. Our study also shows that differences between one-way and two-way coupling approaches are negligible for relatively large sediments, that, on the other hand, are more likely transported according to the bed-load mode. For smaller particles, transported according to the suspension-load mode, the two-way coupling approach reproduces the reduction of turbulence activity already observed in physical experiments.