Identification of carrier- and particle-phase statistical patterns in particle-laden turbulent channel flow using a two-fluid model

The object of this study is particle-laden turbulent flow in a vertical channel, focusing on the mean velocities and longitudinal turbulent velocity fluctuations of both phases. The problem is to predict these characteristics using a two-fluid turbulence model that represents turbulent momentum transfer through the relative motion of two conditional fluids without an additional eddy-viscosity closure. The present study quantitatively validates this formulation for two benchmark conditions against experimental data and V2F results. The specific contribution of the present study is the quantitative validation of this formulation for simultaneously predicting the mean velocities and longitudinal turbulent velocity fluctuations of the carrier and dispersed phases under two benchmark particle-loading conditions. Numerical calculations were performed for 50 μm glass particles St = 0.57, ψ = 0.02 and 70 μm copper particles St = 3.0, ψ = 0.10, and the predictions were compared with published experimental data and V2F results. For the dispersed-phase mean velocity, the two-fluid model gave mean relative errors of 2.85% and 13.08% for the glass- and copper-particle cases, respectively, compared with 3.93% and 33.99% for V2F. For the dispersed-phase longitudinal turbulent velocity fluctuations, the corresponding mean relative errors were 15.49% and 12.29% for the two-fluid model and 24.34% and 18.51% for V2F. The model also reproduced the general shape and near-wall maximum of the carrier-phase turbulent velocity fluctuation profiles, although neither model was uniformly more accurate for this quantity. Appreciable local discrepancies remained in the dispersed-phase turbulent fluctuations for the copper-particle case, particularly in the intermediate wall-normal region. These results demonstrate that the two-fluid formulation can reproduce the principal mean and fluctuating velocity characteristics under the two benchmark conditions considered, while its quantitative accuracy depends on the predicted quantity and wall-normal region. The present validation is limited to these benchmark conditions and does not independently isolate the effects of particle inertia or mass loading. The obtained results are relevant to the numerical modelling of wall-bounded gas–particle flows encountered in pneumatic conveying and particle-separation equipment, where prediction of the mean phase velocities and turbulent particle transport is required

Hali tarjima qilinmagan