Droplet size distribution and mixing hydrodynamics in a liquid-liquid stirred tank by CFD modeling

Stirred tanks are the common unit operations employed for liquid-liquid dispersion in many industries. In this study, computational fluid dynamic (CFD) coupled with population balance modeling (PBM) were applied to simulate mixing of water in crude oils in a stirred tank equipped with Rushton turbine. An Eulerian multiphase model and standard k-epsilon turbulence model were employed to simulate the flow field inside the tank. Experimental results were then used to validate simulation results. The effect of impeller speed, volume fraction of dispersed phase, and oil viscosity on droplet size distribution were investigated in this work. An increase in agitation speed was found to decrease the mean and Sauter mean diameter while increasing the homogeneity of the system. A wider droplet size distribution profile was achieved at higher oil volume fractions with almost no significant change in droplet size. Additionally, increasing the viscosity of the oil phase resulted in a gradual shift towards smaller droplets and narrower droplet size distribution. Liquid-liquid dispersions were occurred in two different turbulent regimes depending on the viscosity of continuous phase. Related correlations for each regime were obtained and the values of maximum droplet size were fitted with the dimensionless numbers. (C) 2019 Published by Elsevier Ltd.