CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Bubble size distribution : Bubble column (

oj.bulmer July 26, 2012 11:07

Bubble size distribution : Bubble column

I was wondering what would be the right way to specify the bubble diameter distribution in CFX for polydispersed phase in bubble column.

1) Polydispersed phase gas with 20 diameter groups
2) Eulerian Inhomogeneous (separate velocity field for liquid and gas), MUSIG homogeneous (One velocity field for all bubble groups)

I have observed that the bubble sizes peak at 6-7 mm and would want to specify fine distribution between 1 to 10 mm while coarse distribution between 10 to 40 mm.

I realise some sort of geometric progression is a right choice. But have no idea what should be exact geometric ratio.

Any help in this would be appreciated.


ghorrocks July 26, 2012 19:34

The MUSIG approaches are designed for this so are likely to be more computationally manageable. But if you have a chance I would try both approaches and see how they compare.

oj.bulmer July 26, 2012 21:30

Am not sure I understood your statement :)

I was talking about only one approach - homogeneous MUSIG in non-homogeneous Eulerian-Eulerian framework. So velocity field for gas and liquid are different. But velocity field for all gas bubble groups is same.

Ideally, the volume averaged size fractions plotted for different bubble diameter groups should be dome shaped and not top heavy, requiring that sufficient range of diameters should be covered to keep the peak somewhere in the middle. This range for my case is 1 to 40 mm as I have deduced.

Now, I need to make sure that I have enough no. of diameter groups in that range, such that any further increase in that number doesn't have significant effect on gas holdup or inter-facial area density. This is like discretization in bubble size group diameters to achieve size group independence.

A geometric progression should be a good fit here. Yet instead of going by a hunch in guessing the geometric ratio, I thought of running it by this forum to see if anyone has stumbled upon a more scientific method.


All times are GMT -4. The time now is 02:05.