Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

A 2D Bubble Column Model

In a bubble column, the gas hold-up is larger in the middle. Rising bubbles induce a circulation pattern, dragging along the liquid in the middle, which circulates downward at the wall. We will model this by using radial profiles of hold-up and velocities. In most of the column, these profiles are constant along the height. However, at the bottom and top of the column, the flow will also be radially directed. These end zones will be modeled as ideally mixed, as indicated in the figure below.

Bubble column schematic

Schematic representation of the bubble-column model with ideally mixed end zones.

Radial profile of gas hold-up

Radial profile of gas hold-up.

Radial profile of liquid velocity

Radial profile of liquid velocity.

Radial profile of turbulent diffusivity

Radial profile of turbulent diffusivity.

The profiles plotted on the right are measured by means of radioactive particle tracking and taken from Degaleesan et al. (Ind. Eng. Chem. Res. 1997, 36, 4670-4680). The column diameter is 19 cm19~\mathrm{cm}, and the height is 190 cm190~\mathrm{cm}. The flow is in the churn turbulent regime.

From the experimental data plotted in the graphs above, we approximate:

εG(r)=0.250.18rR,εL=1εG,vL(r)=vL,0(1r0.7R),Dax=200 cm2 s1+(300 cm2 s1)rR,Drad=40 cm2 s1.\begin{align*} \varepsilon_{G}(r) &= 0.25 - 0.18 \frac{r}{R}, \quad \varepsilon_{L} = 1 - \varepsilon_{G}, \\ v_{L}(r) &= v_{L,0} \left( 1 - \frac{r}{0.7 R} \right), \\ D_{ax} &= 200~\mathrm{cm^{2}~s^{-1}} + \left( 300~\mathrm{cm^{2}~s^{-1}} \right) \frac{r}{R}, \\ D_{rad} &= 40~\mathrm{cm^{2}~s^{-1}}. \end{align*}

We assume that there is a difference ΔvGL\Delta v_{GL} between the gas velocity and liquid velocity, independent of yy, such that:

vG(r)=vL(r)+ΔvGL.v_{G}(r) = v_{L}(r) + \Delta v_{GL}.

The superficial velocity of the liquid equals UL=1 cm s1U_{L} = 1~\mathrm{cm~s^{-1}}, and for the gas: UG=10 cm s1U_{G} = 10~\mathrm{cm~s^{-1}}.

The total height of 190 cm190~\mathrm{cm} will be divided into three parts: a bottom section of 19 cm19~\mathrm{cm}, a middle section of 152 cm152~\mathrm{cm}, and a top section of 19 cm19~\mathrm{cm}. The middle section is modeled by means of the convection-diffusion-reaction equation:

t(εLcL)+x(εLvLcLεLDaxcLx)1rr(r[εLDradcLr])=kLa(cLcL)+εLRL,t(εGcG)+x(εGvGcGεGDaxcGx)1rr(r[εGDradcGr])=kLa(cLcL)+εGRG.\begin{align*} \frac{\partial}{\partial t} \left( \varepsilon_{L} c_{L} \right) + \frac{\partial}{\partial x} \left( \varepsilon_{L} v_{L} c_{L} - \varepsilon_{L} D_{ax} \frac{\partial c_{L}}{\partial x} \right) - \frac{1}{r} \frac{\partial}{\partial r} \left( r \left[ \varepsilon_{L} D_{rad} \frac{\partial c_{L}}{\partial r} \right] \right) &= k_{L} a \left( c_{L}^{*} - c_{L} \right) + \varepsilon_{L} R_{L}, \\ \frac{\partial}{\partial t} \left( \varepsilon_{G} c_{G} \right) + \frac{\partial}{\partial x} \left( \varepsilon_{G} v_{G} c_{G} - \varepsilon_{G} D_{ax} \frac{\partial c_{G}}{\partial x} \right) - \frac{1}{r} \frac{\partial}{\partial r} \left( r \left[ \varepsilon_{G} D_{rad} \frac{\partial c_{G}}{\partial r} \right] \right) &= -k_{L} a \left( c_{L}^{*} - c_{L} \right) + \varepsilon_{G} R_{G}. \end{align*}

Note that the turbulent diffusion coefficients for both liquid and gas are assumed to be identical. This is because small bubbles are dragged along with the liquid.

The top and bottom sections are modeled as ideally mixed. It is advised to model this by using artificially large axial and radial turbulent diffusion coefficients DaxD_{ax} and DradD_{rad} in these sections. Also, ensure that the inflow in the bottom section is actually inflow (i.e., has no negative velocities) and that at the top, there is only outflow.

Questions:

  1. Compute the average hold-ups and determine the values of vL,0v_{L,0} and ΔvGL\Delta v_{GL} such that the prescribed superficial velocities are obeyed.

  2. Implement the formulated model equations in Python for the case of insoluble species (kLa=0k_{L} a = 0).

  3. Model a first-order reaction in the liquid phase. Hint: Because you consider first-order kinetics, the result will be independent of the inlet concentration. A simple choice is cL,in=1 mol m3c_{L,\mathrm{in}} = 1~\mathrm{mol~m^{-3}}. How do the results for a range of kinetic coefficients compare to the conversion in an ideal CSTR and in a plug flow reactor?

  4. Model a first-order reaction in the gas phase (for a range of kinetic coefficients). How does it compare to the conversion in an ideal CSTR and in a plug flow reactor?

  5. Play with the model. Implement a realistic value of kLak_{L} a, a proper equilibrium relation between gas and liquid concentrations (i.e., cL(cG)c_{L}^{*}(c_{G})), and interesting kinetics.

Note that all numerical implementations should be consistent with the provided equations and parameter values.