Many studies on solid-liquid mixing have been dedicated to low density particles at low solids concentrations. In the present work, computational fluid dynamics (CFD) simulation and experimental methods were employed to study suspension of high density particles (nickel) at high solids concentration in water. The work first focused on establishing the velocity field in a liquid-only system and then progressed to a solid-liquid system. In the liquid-only system, the influence of tank geometry and simulation strategies, including turbulence models, on fluid flow pattern and mixing was investigated in a tank stirred by a Rushton turbine. The standard k-f. model gave better overall predictions of mean velocity fields than the k-ro and RNG k-f. models. The CFD simulation and experimental results obtained with the laser Doppler velocimetry (LDV) method showed that mixing time and homogenization energy decreased with a decrease in the impeller bottom clearance. It was further shown that there is a bottom clearance range in which a draft tube can aid mixing in a tank stirred by the Rushton turbine. In the solid-liquid system, a hydrofoil impeller was used to investigate the influence of simulation strategies, particle properties and hydrodynamic operating conditions on mixing features such as the off-bottom solids suspension, cloud height, solids concentration distribution and local particle size distribution. The simulation results were compared with experimental ones, in which the off-bottom solids suspension was determined visually and an optical attenuation technique was employed to determine the cloud height and solids concentration distribution. The local particle size distribution (PSD) in the tank was measured by a laser diffraction method. A better agreement between the simulation and experimental results was obtained with drag models that account for the solids loading or free stream turbulence than those that do not. It was shown that the Stokes law applies up to a diameter of 150 ~m for the nickel particles. A CFD simulation strategy for studying mixing of high density solids is proposed and it is shown that a CFD simulation method can be used to develop empirical models that predict mixing features. A CFD simulation approach that takes particle size into account gives predictions that are more representative of practical applications than the mono-size particle simulation approach. Reactor configurations and hydrodynamic parameters that improve mixing were identified. These can also aid optimal design of mixing systems.
|Number of pages
|Published - 2005