Numerical simulation of a high viscosity bubble column

Abstract

The objective of this work is to develop fluid dynamic model to simulate a high viscosity bubble column for CO2 absorption in Ionic Liquids (ILs). A very promising solvent for CO2 capture and conversion are ionic liquids (ILs); ILs consist of a wide group of salts, which are liquids at room temperature, have low vapor pressure, high ionic conductivity and thermal stability. However, the use of ILs for industrial CO2 depletion has a series of technical and economic issues that must be solved if this strategy is to be implemented. A very important drawback of ILs used for gas removal is its high viscosity, reaching values above 0.010 Pa·s which results in a decrease of the overall mass transfer rate and an increase in the power required for pumping and mixing. In order to elucidate the hydrodynamic behavior in a bubble column for CO2 absorption with one gas feed inlet, a Computational Fluid Dynamic (CFD) model was developed, which was experimentally validated through a laboratory scale bubble column. To simplify the calculations and increase the accuracy of the results, the system was modeled as a single rising bubble which permits the estimation of the bubble rising velocity and the change of the bubble shape and size during its displacement. The model approach consists in a simplified two-dimensional multiphase flow model which considers the liquid solvent as a Newtonian fluid. The laminar, isothermal, and non-stationary hypotheses for both phases is applied. To model the displacement of the gas-liquid interface, the Level Set method was used. The laboratory tests were carried out using water-glycerol mixtures (58 %, 78 %, 84 % and 88 % by weight) and two Imidazolium type ionic liquids (pure [bmim]BF4 and [bmim]PF6). To compare the results obtained from the laboratory and the simulations, the drag coefficient for gas bubbles in liquids was used which correlates the fluid physical properties of fluids and the bubble equivalent diameter and terminal velocity. The results were also compared with predicted values obtained through a new correlation for the drag coefficient of single rising bubbles in ILs proposed by Dong et al. (2010). The results indicated that the CFD model is in good agreement with the experimental results, particularly for bubble Reynolds numbers below 5. Above this value, the model tends to underestimate the bubble terminal velocity which can be explained by the effect of the high velocity gradients close to the gas-liquid interface. Future steps will involve improving of the computational mesh, a parametric analysis of the reintialization parameter and the parameter controlling the thickness at the interface transition zone. Acknowledgments. This work was supported by FONDECYT postdoc N°3120138 from CONICYT (Chile).