Stochastic Lagrangian wall deposition model for RANS prediction of deposition in turbulent gas–solid flows

Abstract

In industrial-scale Reynolds Averaged Navier–Stokes (RANS) simulations, a wall-modeled approach is often employed, specifically choosing a dimensionless wall-cell height much greater than one. However, in gas–solid flows, a coarse description of the boundary layer may lead to inaccurate results in terms of particle dispersion and deposition. In this study, we conducted an investigation on stochastic RANS dispersion models with deposition on smooth walls using a wall-modeled approach. These models were applied to turbulent channel flows in both vertical and horizontal orientations. The modeling approach involved the adoption of the Euler–Lagrange formalism and the RANS $k-ϵ$ turbulence model in OpenFOAM®. To assess particle behavior first, a simple homogeneous isotropic stationary turbulence (HIST) case was examined. Fluid elements were also tracked in an inhomogeneous turbulence for dispersion analysis. These tests gave the good formulation for the dispersion model to be used for particle tracking coupled with turbulent channel flow simulation. However, the deposition results in turbulent channel using such stochastic dispersion model in the wall-cell were found to be consistently overestimated across various scenarios for low-inertia particles. To address these limitations arising from the coarse description of the boundary layer, we examined, implemented, and evaluated a Lagrangian stochastic wall deposition model. The activated deposition model demonstrated good agreement with experimental data.