Iranian Journal of Chemical Engineering (IJChE)

Abstract Computational fluid dynamics (CFD) is a powerful numerical tool that is becoming widely used to simulate many processes in the industry. In this work study of the stirred tank with 7 types of concave blade with CFD was presented. In the modeling of the impeller rotation, sliding mesh (SM) technique was used and RNG-k-ε model was selected for turbulence. Power consumption in various speeds in the single phase, mean tangential, radial and axial velocities in various points, effects of disc diameter and thickness and mixing time were investigated. The optimum concave impeller was selected and the effect of tracer feed position and probe location was investigated on it. Results suggested that power consumption is exactly depending on impellers scale and geometry, was in a good agreement with the experimental data and in turbulent flow is relatively independent of Reynolds number. Power number increases with increasing disc diameter for both concave and Rushton and concave´s power is relatively independent on disc thickness but increasing it decreases Rushton´s power. The data revealed that the power number was 2.3±0.3 for blade angle 40° whereas for blade 25°, 50° and 55° respectively 43% lower and 57% and 43% higher.

Abstract Knowledge of mixing time is of fundamental importance for investigation of mixing efficiency in agitation systems. The mixing time obtained by using the correlation and formula in large scale mixing systems was incorrect. Again, the number of available correlations in this scale of mixing systems is limited. To predict the mixing time of stirred tanks with dual impellers commonly used in industry, a third-compartment mixing model was used. The time of homogenization of the charge (mixing time) was calculated from the time dependency of the local concentration of tracer measured at various locations. Experimental data on mixing time were obtained with a conductivity technique. In the present study distribution of tracer in the bulk of the liquid was described by compartment model (CM) as well as for stirred vessel with dual Rushton impellers. As for the model, a good agreement between the experimental data and the calculated values was apparent.

Abstract "> Experimental measurements of overall gas holdup (εg), residence time distribution (RTD) and liquid mixing time (tm) have been carried out in an air-water system in a 17 lit bubble column with an 11 cm diameter, over a wide range of superficial gas velocity (0.14-1.46 ms-1) and height to diameter ratio (1.36-8.84). The bed expansion method was used to obtain holdup values and the change of slope related to εg versus superficial gas velocity (Usg) indicated the transition point from homogeneous to heterogeneous regime at Usg= 0.7-0.9 ms-1. Besides, the experiments illustrated that H/D ratio had no effect on holdup values. The axial dispersion model was used with semi-closed boundary conditions for prediction of RTD, and hence, the axial dispersion coefficients (Dax). Moreover, fitting results of the model and experimental RTD curves achieved from tracer injection method attained the model parameter, Dax , in three H/D ratios of 4.73, 6.36 and 8.84. Results showed that an elevation in H/D ratio, caused a rise in D ax. On the other hand, the mixing time data declined with an increase in Usg; however enhancement of H/D ratio caused an increase in tm.

Abstract The turbulent flow field generated in a baffled stirred tank was computed by large eddy simulation (LED) and the flow field was developed using the Sliding Mesh (SM) approach. In this CFD study, mixing times and power number have been determined for a vessel agitated by a 6-blade Rushton turbine. The predicted results were compared with the published experimental data. The satisfactory results of comparisons indicate the potential usefulness of this approach as a computational tool for designing stirred reactors.