Some issues concerning Large-Eddy Simulation of inertial particle dispersion in turbulent bounded flows

TL;DR

This study evaluates the accuracy of Large Eddy Simulation in predicting inertial particle dispersion in turbulent wall-bounded flows, revealing limitations even with well-resolved LES and emphasizing the need for improved SGS models.

Contribution

The paper provides a comprehensive analysis of LES limitations in modeling inertial particle dispersion and highlights the importance of higher-order moments in SGS closures.

Findings

LES struggles to accurately predict particle accumulation and segregation.

Reynolds number effects influence particle dispersion, especially for heavy particles.

Higher-order velocity moments are crucial for better SGS modeling of particles.

Abstract

The problem of an accurate Eulerian-Lagrangian modeling of inertial particle dispersion in Large Eddy Simulation (LES) of turbulent wall-bounded flows is addressed. We run Direct Numerical Simulation (DNS) for turbulent channel flow at shear Reynolds numbers equal to 150 and 300 and corresponding a-priori and a-posteriori LES on differently coarse grids. We then tracked swarms of different inertia particles and we examined the influence of filtering and of Sub-Grid Scale (SGS) modeling for the fluid phase on particle velocity and concentration statistics. We also focused on how particle preferential segregation is predicted by LES. Results show that even ``well-resolved'' LES is unable to reproduce the physics as demonstrated by DNS, both for particle accumulation at the wall and for particle preferential segregation. Inaccurate prediction is observed for the entire range of particles considered in this study, even when the particle response time is much larger than the flow timescales not resolved in LES. Both a-priori and a-posteriori tests indicate that recovering the level of fluid and particle velocity fluctuations is not enough to have accurate prediction of near-wall accumulation and local segregation. This may suggest that reintroducing the correct amount of higher-order moments of the velocity fluctuations is also a key point for SGS closure models for the particle equation. Another important issue is the presence of possible flow Reynolds number effects on particle dispersion. Our results show that, in small Reynolds number turbulence and in the case of heavy particles, the shear fluid velocity is a suitable scaling parameter to quantify these effects.