LES mesh resolution requirements for particledriven gravity currents

In this study we investigate the influence of grid resolution on a near-wall resolved LES model of a lock-exchange particle-driven gravity current. The simulations are performed using the finite volume Boussinesq code SnS with a Smagorinsky turbulence model for a buoyant Reynolds number of 60,000 on...

Full description

Bibliographic Details
Main Authors: Pelmard Joë, Friedrich Heide, Norris Stuart
Format: Article
Language:English
Published: EDP Sciences 2018-01-01
Series:E3S Web of Conferences
Online Access:https://doi.org/10.1051/e3sconf/20184005028
Description
Summary:In this study we investigate the influence of grid resolution on a near-wall resolved LES model of a lock-exchange particle-driven gravity current. The simulations are performed using the finite volume Boussinesq code SnS with a Smagorinsky turbulence model for a buoyant Reynolds number of 60,000 on 4 grid sizes. According to previous studies, two-point correlations are most appropriate to estimate LES resolution. With the largest scales of the flow being resolved by more than 20 cells, well-resolved LES is obtained for grid resolutions of 1925×62×125 and finer. In addition, in order to apply the turbulence model correctly, we show that the velocity power spectrum densities provide useful information for the maximum cell size. The ratio of the subgrid scale viscosity to the molecular viscosity and the subgrid scale shear-stress to the resolved Reynolds stress show good convergence with grid refinement. The ratios vSGS / <0.3 above the current and τSGS (u'v')ave < 0.05 inside the mixing layer, are chosen as threshold values, based on our evaluation study.
ISSN:2267-1242