(324a) Cfd Simulation of Bubble Columns: Modeling of Non-Uniform Gas Distribution at Sparger

Most of the CFD (Computational Fluid Dynamics) simulations of bubble column reactors are based on the assumption that gas is uniformly distributed at the sparger. All the investigators adjust lift and drag coefficients to match overall gas hold-up and liquid circulation (see Joshi, 2001and references cited therein). The experimental data used to evaluate and fit these model parameters is usually obtained with sieve plate sparger. The sieve plate spargers are known to lead to non-uniform gas distribution. Only fraction of total numbers of holes of such a sieve plate sparger are active at any instant and this leads to non-uniform distribution of gas. The active region of the sparger varies with time leading to complex unsteady flow. Typical dynamic flow behavior on the sparger plate for the air water system is shown in Figure 1. It is obvious that it is essential to capture non-uniform distribution of gas at the sparger for realistic simulations of gas-liquid flow in bubble columns. Such an attempt is made here. More than a decade ago, Ranade (1993) have attempted to account for non-uniform distribution of gas and the sparger. However, the model considered axis-symmetric two-dimensional flow in bubble column and was a steady state model. In this work we have developed a comprehensive, three-dimensional, unsteady CFD model to simulate gas-liquid flow in bubble columns after accounting for non-uniform distribution of gas at the sparger. In this work, a bubble column (air-water system) of 0.2 m diameter and 2 m height with the uniform-hole sparger (hole diameter 1.2 mm and F.A. of 1.13%) was used (Figure 2a). The dispersed height to diameter ratio of 5 was maintained with the superficial gas velocities ranging from 0.05-0.40 m/s. Wall pressure fluctuations were measured using dynamic pressure transducers (PCB Piezotronics Inc.,
USA) located at different axial locations from the sparger. CFD model was developed using Eulerian approach for two-phase system. The standard k-e mixture turbulence model was used to incorporate effect of turbulence. Both the plenum section and the sparger plate above were considered in the solution domain for the first time (Figure 2b). The uniform gas velocity was set as the inlet to the bottom of plenum section considering the volume fraction of gas and bubble rise velocity (Rampure et al., 2007). Sparger was modeled as porous zone. The viscous resistance and the inertial resistance for the perforated plate were estimated using the correlation suggested by Smith and Van-Winkle (1958). CFD simulations were carried out using commercial CFD code, FLUENT 6.3 (Ansys-Fluent Inc.,
USA).

Time-averaged global and local hydrodynamic properties from the CFD simulations were validated with the experimental results of Rampure et al., 2007. Dynamic flow characteristic behavior by the sparger was captured quantitatively using wall pressure fluctuations. The effect of resistances in the porous zone (sparger) was studied to estimate the changes in flow characteristics such as gas holdup, liquid velocity profiles and mixing time. Almost uniform distribution of the gas holdup on the sparger plate was observed for lower gas velocity compared to that at higher gas velocity of 0.2 m/s. The dynamic nature for flow on the sparger was more predominant for increase in gas velocity. Increase in sparger resistance (leading to more uniformity in gas distribution) for lower gas velocity lead to increase in the predicted gas holdup. A sample of predicted results is shown in Table1. The approach, models and results presented in this work will be useful for understanding influence of sparger on gas-liquid flow in bubble columns and will lead to more reliable CFD models and parameters for further work.

References

·          Joshi, J. B., (2001). ?Computational flow modeling and design of bubble column reactors?, Chem.
Eng.Sci. 55, 5893.

·          Rampure M. R., A.A. Kulkarni and V.V. Ranade, (2007). ?Hydrodynamics of bubble column reactors at high gas velocity: Experiments and CFD simulations,? accepted,Ind.
Eng. Chem. Res.

·          Smith, P. L. and M. Van Winkle, (1958). ?Discharge coefficients through perforated plates at Reynolds number of 400 to 3000, AIChEJ, 266.