Three-phase Eulerian simulations of bubble column reactors operating in the churn-turbulent regime: a scale up strategy

https://doi.org/10.1016/S0009-2509(99)00582-5Get rights and content

Abstract

This paper develops a strategy for scaling up bubble column reactors operating in the churn-turbulent flow regime using computational fluid dynamics (CFD). The bubble column is considered to be made up of three phases: (1) liquid, (2) “small” bubbles and (3) “large” bubbles and the Eulerian description is used for each of these phases. Interactions between both bubble populations and the liquid are taken into account in terms of momentum exchange, or drag, coefficients, which differ for the “small” and “large” bubbles. The interactions between the large and small bubble phases are ignored. The turbulence in the liquid phase is described using the kε model. The three-phase description of bubble columns was implemented within the Eulerian framework of a commercial code CFX 4.2 of AEA Technology, Harwell, UK. Two types of approaches were first compared: (a) a simulation model assuming axi-symmetry and (b) a complete three-dimensional model for the cylindrical columns. The three-dimensional simulation showed chaotic behaviour. After averaging with respect to time and in the azimuthal direction, the radial distribution of liquid velocities corresponded closely with the two-dimensional axi-symmetric model. The total system gas hold-up predicted by these two simulation variants were also comparable though there was a significant difference in the radial distribution of the hold-up profiles of the large and small bubbles. For purposes of validation of the three-phase Eulerian simulation model, experiments were carried out in columns of 0.1, 0.174, 0.19, 0.38 and 0.63 m diameter. Three types of experiments were carried out: (1) dynamic gas disengagement experiments to determine the hold-ups of small and large bubble populations, (2) radial distribution of the axial component of the liquid velocity, and (3) centre-line liquid velocity. Demineralized water and Tellus oil, with a viscosity 75 times that of water, were used as liquid phase and air as gaseous phase. Comparison of the experimental measurements with the Eulerian simulations was used to conclude that the two-dimensional axi-symmetric model is adequate for scale up purposes. Simulations for columns with diameters ranging from 1 to 6 m were carried out to emphasise the strong influence of scale on the hydrodynamics.

Introduction

Bubble column reactors operated in industry have several distinguishing features: (1) large column diameters are involved, ranging to 6 m, (2) high superficial gas velocities, in the 0.1–0.4 m/s range, are usually used, (3) the system pressure can range to 6 MPa and (4) the liquid phase often consists of a non-aqueous hydrocarbon mixture (Krishna, Ellenberger & Sie, 1996). Laboratory studies on bubble column hydrodynamics are usually carried out with the air-water system, at ambient pressure conditions, in columns that are smaller than say 0.5 m in diameter (Deckwer, 1992). Even for the air–water system, available literature correlations give significantly different results. This is demonstrated by the predictions of the total gas hold-up and the centre-line liquid velocity as a function of the superficial gas velocity and column diameter; see Fig. 1, Fig. 2. Only two correlations plotted in Fig. 1 anticipate that the gas hold-up decreases with increasing column diameter. We see from Fig. 2(b) that the predictions of the centre-line velocity for a bubble column of diameter 6 m diameter operating at U=0.3 m/s varies between 0.9 and 4.5 m/s. This represents a variation of a factor of five and so there is a clear need for a reliable scale up strategy.

The major objective of the present paper is to develop a model for predicting the scale dependence of the hydrodynamics of bubble column reactors operating in the churn-turbulent regime. The model is based on computational fluid dynamics (CFD) and uses an Eulerian description for the fluid phases. We attempt to validate, at least partially, the scale dependence predicted by the CFD model by comparison with experimental data generated in our laboratory in columns ranging in diameter from 0.1 to 0.63 m. Both experimental data from our data bank, partly published previously (Krishna & Ellenberger, 1996; Krishna, Urseanu, van Baten & Ellenberger, 1999b), and new experimental data generated in this work have been used for validation purposes. Furthermore, for purposes of validation of the CFD simulations we also use the experimental data on the radial distribution of gas hold-up and liquid velocity obtained by Hills (1974) in a 0.14 m diameter column with the air–water system.

Section snippets

Experimental

Two types of experiments were performed: (1) Dynamic gas disengagement experiments to determine the hold-ups of the “small” and “large” bubble populations and (2) Measurement of the radial distribution of the axial component of the liquid velocity.

The description of the dynamic gas disengagement experiments data analysis procedure has been discussed earlier (Ellenberger & Krishna, 1994; Krishna & Ellenberger, 1996; Krishna, Van Baten & Ellenberger, 1998). For this study additional measurements

Development of CFD model

For the homogeneous regime of operation of bubble columns a more or less uniform bubble size is obtained (Clift, Grace & Weber, 1978). Many CFD approaches have been successfully developed to cater for this homogeneous regime of operation using the Eulerian description for the gas and liquid phases (Boisson & Malin, 1996; Grevskott, Sannæs, Dudukovic, Hjarbo & Svendsen, 1996; Grienberger & Hofmann, 1992; Jakobsen, Sannæs, Grevskott & Svendsen, 1997; Kumar, Vanderheyden, Devanathan, Padial,

Simulations vs. experiments

We first compare the results of two-dimensional axi-symmetric simulation with a complete three-dimensional simulation of a 0.38 m diameter column at U=0.23 m/s with the air–water system. Fig. 4(a) shows the transient approach to steady state in the 2D simulation. The parameter values at the end of the simulation were taken to be the steady-state values. The corresponding 3D simulation shows chaotic behaviour (cf. Fig. 4(b) and (c)), which can best be appreciated by viewing the animations on our

Conclusions

The following major conclusions can be drawn from the present work.

1. For reasonable predictions of radial distribution of liquid velocity and gas hold-up we must resort to complete three-dimensional Eulerian simulations.

2. For estimation of average gas hold-ups in the dispersion and circulating liquid velocities, typified by the centre-line velocity VL(0), two-dimensional simulations assuming cylindrical axi-symmetry are of adequate accuracy.

3. On the basis of the comparison of Eulerian

Acknowledgements

The Netherlands Organisation for Scientific Research (NWO) is gratefully acknowledged for providing financial assistance to J.M. van Baten.

References (44)

  • R. Krishna et al.

    Influence of scale on the hydrodynamics of bubble columns operating in the churn-turbulent regime: Experiments vs Eulerian simulations

    Chemical Engineering Science

    (1999)
  • A. Lapin et al.

    Numerical simulation of the dynamics of two-phase gas–liquid flows in bubble columns

    Chemical Engineering Science

    (1994)
  • Y. Ohki et al.

    Longitudinal mixing of the liquid phase in bubble columns

    Chemical Engineering Science

    (1970)
  • J. Sanyal et al.

    Numerical simulation of gas–liquid dynamics in cylindrical bubble column reactors

    Chemical Engineering Science

    (1999)
  • A. Sokolichin et al.

    Gas–liquid flow in bubble columns and loop reactors: Part I. Detailed modelling and numerical simulation

    Chemical Engineering Science

    (1994)
  • A. Sokolichin et al.

    Applicability of the standard — turbulence model to the dynamic simulation of bubble columns: Part I. Detailed numerical simulations

    Chemical Engineering Science

    (1999)
  • R. Torvik et al.

    Modelling of slurry reactors. A fundamental approach

    Chemical Engineering Science

    (1990)
  • K. Akita et al.

    Gas hold-up and volumetric mass-transfer coefficient in bubble columns

    Industrial and Engineering Chemistry, Process Design and Development

    (1973)
  • H.F. Bach et al.

    Variation of gas hold-up in bubble columns with physical properties of liquids and operating parameters of columns

    German Chemical Engineering

    (1978)
  • Bernemann, K. (1989). Zur Fluiddynamik und zum Vermischungsverhalten der flüssigen Phase in Blasensäulen mit...
  • N. Boisson et al.

    Numerical prediction of two-phase flow in bubble columns

    International Journal of Numerical Methods in Fluids

    (1996)
  • R. Clift et al.

    Bubbles, drops and particles

    (1978)
  • Cited by (141)

    View all citing articles on Scopus
    View full text