RANS-based predictions of dense solid–liquid suspensions in turbulent stirred tanks

This work is aimed at investigating the impact of different meso-scale models and constitutive equations for the RANS-based two-fluid model simulations of a turbulent solid–liquid stirred vessel with high solids loading. The model assessment is preceded with a grid convergence study, which confirms the variability of the discretization requirements depending on the observed variable. The simulation results demonstrate that for the investigated system the high solids loading contribution modelled by the kinetic theory of granular flows is negligible, both in incomplete and complete suspension conditions. Instead, the particle concentration fluctuations contribution included in the momentum equations dramatically affect the predictions, particularly in incomplete suspension conditions. The evaluation of the models is completed by the comparison of the predicted solids concentration profiles with experimental data measured by Electrical Resistance Tomography. The computational strategy for achieving realistic predictions of the solid distribution both in complete and incomplete suspension conditions is outlined.