Two-phase Flow CFD Simulations And Its Study

Solid particles' dispersion or suspension in a fluid phase is often a practise in many Industrial processes. The particle drag coefficient is one of the most important hydrodynamic parameters involved in the modeling and design of such processes. Many empirical functions and data correlations for predicting the drag coefficient can be found in literatures. For some simple cases, those correlations are good enough to give accurate results. However, they could not be applicables when applied to industrial process. The objective of the work reported in this thesis is to integrating computational fluid dynamics (CFD) and two-phase flow models to predict the drag coefficient. A commercial CFD code FLUENT is employed. In the simulations, the solid settling velocities were taken as the index representing the drag coefficient and compared with the experimental ones. The CFD model built in this work is based on a specially designed experimental set-up by Brucato et. al. for measuring drag coefficient. 3D single phase flow fields were simulated at three rotation speeds (300rpm, 500rpm and 700rpm). Two different turbulence models (k-ε and k-ε RNG model) were used for comparisons. Based on developed single phase, Euler multi-phase model was used to predict the solid settling in a turbulence field. The effects of different drag coefficient empirical expressions inlet velocity on the settling velocities were evaluated. The results show the predicted settling velocities are in significant disagreement with experimental data, which implys the correlations selected in this work are not suitable since Euler model requires a pre-defined drag coefficient correlation. The discete phase in Lagrangian reference frame was introduced to the single flow field predicted. Solid particles settling velocities were estimated under the conditions of three stir speeds (300rpm, 500rpm and 700rpm). The influences of different step length facors, different turbulence models and different injection types on the predictions were also investingated. Comparing with experimental data, Lagrangian multi-phase model gives more accurate predictions for particle settling velocities.