Numerical Simulations of Gas-Liquid Flow Dynamics in Bubble Columns

Abstract

There is great potential for using computational fluid dynamics (CFD) as a tool in scale-up and design of bubble columns. Full-scale experimentation in bubble columns is expensive and CFD is an alternative approach to study bubble column hydrodynamics. However, CFD can be computationally intensive as a predictive tool for a full three-dimensional geometry. In this paper, a 0.2 m diameter semi-batch bubble column is numerically simulated and the results are compared to experimental measurements performed by Rampure et al. [1]. The objectives are to examine and determine an appropriate set of numerical parameters and to determine if two-dimensional simulations are able to accurately predict observed bubble phenomena so that the computational cost can be reduced. A two-fluid Eulerian-Eulerian model is employed to represent each phase as interpenetrating continua and the conservation equations for mass and momentum for each phase are ensemble-averaged. Time-averaged gas holdup is mainly examined due to its significant role in gasliquid mass transfer and to compare to available data. Numerical predictions are presented for gas holdup at various axial heights as a function of radial position for a superficial gas velocity of 0.1 m/s. The numerical predictions exhibit the axial development of the gas holdup profile phenomena; that is, the gas holdup at the center of the column increases with increasing axial height. The effects of grid resolution and convergence criteria on the numerical predictions are also demonstrated. However, FD ca n be computationally inten ive a a predict ive too l fo r a full th reedimensional geo metry. In th i paper, a 0.2 m diameter semibatch bubble column is numerica ll y imul ated and the res ults are compared to ex perimental meas urements performed by R ampure et al. [ !] . The objecti ve are to exa mine and determine an appropriate set of numerica l parameters and to determine if two-dimensional simul ations are abl e to accuratel y predtct observed bubble phenomena so that the computational cost can be reduced. A two-fluid Euleri an-Eul eri an model is employed to represent each phase as interpenetrating continua and the conservation equations fo r mass and momentum for each phase are ensemble-averaged. Time-averaged gas holdup ts mamly examined due to its significant role in gas-liquid mass transfer and to compare to available data. Numerical predictions are presented for gas holdup at various ax ial heights as a function of radial position for a superfi cial gas velocity of ~· 1 mls. The numerical predictions exhibit the axial evelopment of the gas holdup profile phenomena · that is the gas h ld ' '. 0 up at the center of the column increases with Increasing axial height. The effects of grid resolution and convergence .t . d en ena on the numerical predictions are also emonstrated.

Stability Issues for Gas-Liquid Flows in Bubble Columns

“…The number of computational cells along the horizontal and vertical directions is increased while maintaining a square computational cell; a summary of grid resolutions is presented in Table 1. The effects of grid resolution on the numerical predictions of the time-averaged gas holdup were examined and compared to previous results by the authors using FLUENT [2].…”

“…Figure 2 shows the predictions of the time-averaged gas holdup using FLUENT for axial locations of 0.15 m and 0.65 m above the distributor plate [2].…”

“…Law et al [2] studied bubble column hydrodynamics computationally using the Eulerian-Eulerian method formulated in FLUENT. Unphysical results were observed with increasing grid resolutions using FLUENT, therefore prohibiting a conclusive grid resolution study.…”

“…The bubble pressure model is reported to account for bubble stability, thus providing physical solutions. The objectives are to obtain grid-independent numerical solutions to resolve unphysical results observed in FLUENT with increasing grid resolutions [2], and to validate computational fluid dynamics simulations with experimental data to demonstrate the use of numerical simulations as a viable design tool for gas-liquid bubble column flows. …”

“…The objectives are to obtain grid-independent numerical solutions to resolve unphysical results observed in FLUENT with increasing grid resolutions [2], and to validate computational fluid dynamics simulations with experimental data to demonstrate the use of numerical simulations as a viable design tool for gas-liquid bubble column flows . turbulent Schmidt number for k and e , respectively effective stress (N/m 2 ) interactions, and mixing rate [ 4,5]. Full-scale experimentation in bubble columns is expensive; a more cost-effective approach to exploring these reactors is by using validated computational fluid dynamics (CFD) models.…”