LES simulations of sheet flow sediment transport

A great diversity of complex physical processes contribute to the transport of natural sediments by environmental water flows. Typically, the particles can be carried out in suspension (suspended-load), i.e without contact with the bottom, or near the bed (bedload), i.e in permanent or intermittent contact with the flow bed by rolling, bouncing, jumping or saltating. In the suspension, the sediment concentration is low and the particle motion is driven by water turbulence. Near the bed, the concentration of the mobile sediments increases significantly and reaches maximal values between 50 % and 60 % in volume. In this near bed region, particle-particle interactions (collisions, friction) are the dominant mechanism for momentum transfer, as fluid turbulence is damped due to the presence of particles.

Modelling sediment transport is still a challenge and most sediment transport models are based on the “passive scalar” hypothesis for the suspension while the bed load is modelled as an empirical sediment flux. Such a splitting of the domain into two subdomains that can exchange sediment is somehow arbitrary and does not represent the existence of a continuum over the water column.

This ambiguity is a strong motivation for the development of two-phase flow models in which the motion of both the particles and water are solved, and this project proposes a complementary experimental, modelling and simulation approach to solve this problem, relying on recent experimental and numerical results.

The idea is to perform two-phase turbulent resolving simulations (LES) of intense bed-load sediment transport (sheet flow) for which high-resolution experimental data have recently been obtained at the LEGI. After a careful validation, the Double Averaging Methodology is applied to the LES simulations to provide more insight into the complex couplings between the boundary layer turbulent macro-structures and the particles dynamics.

The two experimental observations that we want to address are:

• the strong damping of turbulence illustrated by the reduction of the von Karman constant by a factor of two,
• the increase of concentration diffusivity illustrated by a Schmidt number of about 0.4.

The numerical simulations should provide the missing spatial resolution (stream-wise and span-wise directions) of the experimental data to further identify the mechanisms responsible for these two significant turbulence modifications.

In a second step, the role of particle inertia on these mechanisms will be addressed by using different sediment types (size and density) for which new experimental data will be acquired during the Hydralab project, H2020.

The von Karman reduction has been observed for decades, but no satisfactory explanations have been given so far. Providing some insights on this observation would be a scientific breakthrough for both the sediment transport and the multi-phase flow communities.