Quantification of Numerical and Model Uncertainties in the Cfd Simulation of the Gas Holdup and Flow Dynamics in a Laboratory Scale Rushton-turbine Flotation Tank

The gas holdup and flow dynamics in a laboratory scale Rushton-turbine flotation tank were simulated using the dispersed k- turbulence model and Eulerian-Eulerian multiphase modelling approach. The inherent uncertainties in the CFD predictions were quantified by examining the model and numerical (i.e., discretization) uncertainties. The contribution of model uncertainty was explored by considering different sparger geometries while numerical uncertainty was evaluated using the Grid Convergence Index (GCI). The numerical predictions of the gas holdup for different sparger configurations showed that sparger design and location can have a significant impact on the results. Numerical uncertainty was found to be less significant since the average GCI value was less than 11.7%, for the numerical predictions of the velocity components and the turbulence parameters. Overall, the results suggest that quantification of uncertainties in CFD simulations can lead to improved predictions of the gas holdup and flow dynamics in a laboratory scale Rushtonturbine tank.