Gas holdup and hydrodynamic flow regime transition in bubble columns

https://doi.org/10.1016/j.jiec.2017.07.043Get rights and content

Highlights

  • The effects of gas density, surface tension, and viscosity on gas holdup and flow regime transition were studied.

  • The dual effect of viscosity was observed and analyzed.

  • Transition gas holdup was correlated as a function of the gas density, surface tension, and liquid viscosity.

Abstract

The homogeneous-to-heterogeneous flow regime transition point dependence on gas and liquid properties was investigated in a semi-cylindrical bubble column of 1.8 m height and 0.21 m inner diameter operating as a semi-batch system. He, air, and CO2 gases were injected at superficial gas velocities of up to 239 mm/s. The batch liquids included water, aqueous ethanol solutions, and aqueous glycerol solutions, all with a gas-free liquid height settled at 1 m. When the gas density increased, the gas holdup increased at all superficial gas velocities, delaying the flow regime transition. The gas holdups in the liquid mixtures were higher than those for tap water. The transition gas holdup for the ethanol solutions increased to a sharp maximum and then decreased as the surface tension increased. Also, the glycerol solutions showed similar behavior with respect to increasing liquid viscosity, but with a shallower maximum. The transition gas holdup was empirically correlated as a function of the gas density, surface tension, and liquid viscosity, employing dimensional constants. The measured transition gas holdups for liquid mixtures, as well as some data from the literature, were fitted by the correlation.

Introduction

Bubble columns provide a classical operational method for contacting liquids with continuous gas flow, and are widely used in chemical, biochemical, and petrochemical processes. Gas holdup control in bubble columns is critical in reactor design and modeling.

In a solids-free bubble column, two main flow regimes are observed — homogeneous and heterogeneous. The homogeneous regime is characterized by relatively small uniform gas bubbles, and the gas holdup increases almost linearly with increasing superficial gas velocity. On the other hand, the heterogeneous regime or churn-turbulent regime is characterized by vigorous bubble coalescence and break-up, high bubble rise velocities, and much larger, less-uniform bubbles. When the homogeneous regime undergoes transition to the heterogeneous regime as the superficial gas velocity is increased, the slope of the gas holdup vs. gas velocity changes substantially [1]. Demarcation of the transition between the homogeneous and heterogeneous flow regimes is essential for operators of bubble columns.

According to Zahradnik et al. [2] and Ruzicka et al. [3], the flow regime transition region clearly depends on the orifice diameter and open area fraction of the gas distributor. After comparing flow regimes from the distributors that have hole sizes of 0.5 and 1.6 mm, for a fixed open ratio of 0.5%, Zahradnik et al. [2] reported that the flow regime transition and heterogeneous regime are difficult to characterize. Camarasa et al. [4] identified the change of flow regime when using a porous plate, as well as for single- and multiple-orifice nozzles. For the porous plate and multiple-orifice nozzle distributor, the gas holdup increased and then decreased with increasing superficial gas velocity in the range of the flow regime transition. Kazakis et al. [5] classified the flow in this region into three regimes: pseudo-homogeneous regime, transition regime, and heterogeneous regime. They considered the pseudo-homogeneous regime as the case where a linear increase of gas holdup occurs with increasing, yet low, superficial gas velocity. Finch and Dobby [6] and Bennet et al. [7] identified the transition as the point where the slope of the gas holdup vs. superficial gas velocity changes abruptly.

The drift flux model by Wallis [8] is commonly used to identify the flow regime transition. This model introduces the critical point, which initiates the flow regime transition when the stability of the homogeneous regime begins to diminish. In the range of superficial gas velocities in the homogeneous regime, the data show a negative parabolic form when the drift flux (explained in Section Drift flux theory) is plotted vs. the superficial gas velocity. When the bubble stability diminishes, the drift flux no longer increases linearly with gas holdup.

The flow regime transition depends on various factors: the gas distributor type, gas density, liquid viscosity, and surface tension. Numerous empirical correlations for gas holdup have been proposed which do not distinguish between the separate homogeneous and heterogeneous regimes [9], [10], [11], [12].

A number of research groups have determined the flow regime transition experimentally. Table 1 shows the operating and design conditions for various previous studies on the flow regime transition. The empirical correlations by Reilly et al. [13] and Wilkinson et al. [14] are the most widely used flow regime transition correlations for bubble column reactors. However, these equations do not match the data for liquid mixtures, especially at low solute concentrations. Moreover, liquid viscosity was not considered in the flow regime transition gas holdup correlation by Reilly et al. [13], and the equation by Wilkinson et al. [14] does not provide a good fit to data that show the effect of liquid viscosity.

In the current study, in order to improve the ability to predict the gas holdup at the flow regime transition, we varied the liquid surface tension and viscosity, using aqueous ethanol and glycerol solutions. In addition, to investigate the effect of gas density, several different gases were tested. An empirical correlation is proposed which considers these properties.

Section snippets

Experimental

Fig. 1 shows a schematic of the experimental equipment. Gas holdup and axial pressure drop estimation were carried out with a semi-cylindrical acrylic column of a 1.8 m height and 0.21 m inner diameter. The column was filled with the water or aqueous solutions up to a 1 m height and the liquid temperature was fixed at 20 ± 2 °C. The upper surface of the liquid in the column was exposed to the atmosphere. A gas distributor was positioned at the bottom of the test section to distribute the gas flow

Drift flux theory

In the homogeneous regime, a stable uniform bubble size leads to a linear increase in gas holdup with increasing superficial gas velocity. The point at which the gas holdup variation departs from this linear relationship is the flow regime transition point. Bennet et al. [7] identified this flow regime transition point by graphically plotting the gas holdup against the gas flow rate. However, Olivieri et al. [15] stated that a plot of εg vs. Ug cannot clearly distinguish the boundary of the

Influence of gas density

Fig. 4 shows the dependence of the gas holdup on the gas density and gas velocity for the gas–water system at different gas velocities. When the gas density increases, the gas holdup increases due to the longer residence time of the bubbles in the column. Hecht et al. [18] also found that high density gases produce higher gas holdup even when the bubble sizes are the same because the buoyancy and interfacial effects (i.e., surface tension) are affected by the gas density. At low superficial gas

Empirical correlation

Several researchers [13], [14], [39], [40] have proposed empirical or semi-empirical correlations of gas holdup and superficial gas velocity at the flow regime transition point. Of the researchers mentioned above, only Wilkinson et al. [14] and Reilly et al. [13] have suggested empirical correlations of gas holdup at the transition point. However, these two empirical correlations, as mentioned above, do not fit our experimental data. This is because the transition gas holdup of the liquid

Conclusions

The flow regime transition gas holdup was determined as a function of the gas density, liquid viscosity, and surface tension in a stationary liquid bubble column with 2 mm sparger openings. When the gas density increased, the gas holdup increased at all superficial gas velocities above 50 mm/s tested. Gas holdup at the transition point also increased due to the stabilization of bubbles by increased gas density. For dilute aqueous ethanol solutions, the gas holdup was greater than that for tap

After-word

The maxima observed in Fig. 7, Fig. 9 for liquid surface tension, and Fig. 10, Fig. 12 for liquid viscosity, as has been pointed out by others, e.g. Wilkinson et al. [14] for viscosity, differ from what is observed for monocomponent liquids, which all show unidirectionally downward curves over the same liquid property ranges. The dynamic surface tension of the ethanol solutions probably differ from their measured static or equilibrium values and, especially in the case of the glycerol

Acknowledgement

We gratefully acknowledge the financial support provided by the R&D Convergence Program of the Ministry of Science, ICT and Future Planning (MSIP) and the National Research Council of Science & Technology (NST) of the Republic of Korea (CRC-14-1-KRICT). This work was financially supported by a grant from the National Research Foundation (NRF) of Korea (NRF-2016M3D3A1A01913253).

References (46)

  • J. Zahradnik et al.

    Chem. Eng. Sci.

    (1997)
  • M.C. Ruzicka et al.

    Chem. Eng. Sci.

    (2001)
  • E. Camarasa et al.

    Chem. Eng. Process.

    (1999)
  • N.A. Kazakis et al.

    Chem. Eng. Sci.

    (2007)
  • J.A. Finch et al.

    Int. J. Mineral Process.

    (1991)
  • G. Olivieri et al.

    Chem. Eng. Sci.

    (2011)
  • A. Shaikh et al.

    Flow Meas. Instrum.

    (2005)
  • R. Krishna et al.

    Chem. Eng. Sci.

    (1996)
  • H.M. Letzel et al.

    Chem. Eng. Sci.

    (1997)
  • G. Besagni et al.

    Chem. Eng. Sci.

    (2017)
  • U.P. Veera et al.

    Chem. Eng. J.

    (2001)
  • H. Chaumat et al.

    Chem. Eng. Sci.

    (2007)
  • R. Krishna et al.

    Int. Commun. Heat Mass Transfer

    (1999)
  • S.H. Eissa et al.

    Chem. Eng. Sci.

    (1975)
  • M.C. Ruzicka et al.

    Chem. Eng. J.

    (2003)
  • G. Kuncova et al.

    Chem. Eng. Process.

    (1995)
  • J.J. Heijnen et al.

    Chem. Eng. J.

    (1984)
  • G. Besagni et al.

    Chem. Eng. Sci.

    (2017)
  • R. Krishna et al.

    AIChE J.

    (1994)
  • M.A. Bennet et al.

    Chem. Eng. Sci.

    (1999)
  • G.B. Wallis

    One Dimensional Two-Phase Flow

    (1969)
  • K. Akita et al.

    Ind. Eng. Chem. Process Des. Dev.

    (1973)
  • A. Kumar et al.

    Can. J. Chem. Eng.

    (1976)
  • Cited by (32)

    View all citing articles on Scopus
    View full text