Stochastic modeling for subgrid-scale particle dispersion in large-eddy simulation of inhomogeneous turbulence
Maria Knorps, Jacek Pozorski
Abstract
We consider Lagrangian modeling of heavy particle motion in inhomogeneous turbulence. The dynamics of point particles is one-way coupled to the large-eddy simulation (LES) of fluid flow. To account for the effect of non-resolved (subgrid) flow scales on particle motion, we propose a model for the fluid velocity along the particle trajectories. The model, based on a stochastic diffusion process, accounts for turbulence anisotropy and utilizes the statistical estimates of subgrid quantities: the velocity components (the r.m.s. and covariance) and the Lagrangian time scales. The turbulent channel flow case is taken for validation. First, we discuss the outcome of an a priori LES study. Then, the proposed subgrid dispersion model is tested in a true LES computation. The resulting velocity statistics, particle concentration profiles, and the deposition velocity are compared against available reference data from direct numerical simulations.