[Xilaf]

Hi,

I simulate air bubble rise velocity in water column, in Fluent. I have a serious problem with the results:

A small bubble (d=1 mm) rises in the simulation cca 30 mm/s. It is too slow. It should rise 200mm/s, how can I see from literature (R.Clift at al,1978) and from experiment with camera.
A bubble d=2 mm rises cca 200 mm/s. It should rise 300mm/s.

In the simulation I use the VOF model. Is it right?
Do you know, where could be a problem?

I attached the .jpg file with the situation at the beginning of the simulation.

Thank you

[sega]

Did you try the axi-symmetric destcription of the domain?
I think in your case you are simulating infinite cylinders instead of bubbles.

[thecfduser]

give us more specifications...
Are u solving in 3D??? did u try to refine the mesh???? the value of Clift had it been done in contaminated or very clean water ????

[nikhiljain.iitk]

Originally Posted by Xilaf

Hi,

I simulate air bubble rise velocity in water column, in Fluent. I have a serious problem with the results:

A small bubble (d=1 mm) rises in the simulation cca 30 mm/s. It is too slow. It should rise 200mm/s, how can I see from literature (R.Clift at al,1978) and from experiment with camera.
A bubble d=2 mm rises cca 200 mm/s. It should rise 300mm/s.

In the simulation I use the VOF model. Is it right?
Do you know, where could be a problem?

I attached the .jpg file with the situation at the beginning of the simulation.

Thank you

Yes VOF model is the right model for you ...
but the point is are you using Correct Rise Velocity for bubbles?
if you want more specific details I can arrange.....

[Xilaf]

I'm not using a rice velocity. I only want to set the parametres and physical properties of fluids (bubble, water) at the beginning of the bubble rising and from the results read the right rise velocity of the bubble.

But in VOF I can't set the drag coefficient.

I use the Fluent database fluids (clean water in 20 °C).

I'm solving in 2D. I didn't try to refine the mesh. I work with mesh consists from quadrilaterals (0.1 x 0.1 mm). The smallest bubble I solve is 0.5 mm in diameter.

If you could tell me more specific details, I will be glad.
Thanks

[Xilaf]

I haven't tried the axi-symmetric geometry in VOF model. Only a planar geometry. Do you think, it is wrong and the axi-symmetric geometry can help me to get better results?

[sega]

Originally Posted by Xilaf

Do you think, it is wrong and the axi-symmetric geometry can help me to get better results?

Yes, I made that experience.

[Philipov]

define max height of the cylinder ... it can not be infinite..... the static pressure above the bubble plays role in the velocity reduction... if you compare the values to an experiment then create the same experimental model - diameter, water properties (if water is used) and so on.... if you deal with VOF... need to specify the interaction between phases... this may be you did not do....

[Xilaf]

The infinite cylinder doesn't mean, that the water column is infinity, but that the circle midhip doesn't represent the bubble but the infinite cylinder orthogonal on the 2D geometry.

I set the values of bubble (diameter) and both phases (density, viskosity). and the interaction between phases well, I think. (0.073 mN/m)

[sega]

You have been asking for days or maybe weeks now.
What don't you give it a try and set up an axi-symmetric case?
This can't take longer than you have wasted on asking the same question over again.

[Xilaf]

I understand you. I'm going to try it, but I haven't had the approach to Fluent few days. So now I can only communicate here or study the theory...

[thecfduser]

u did a very big mistake my friend.....
Ure problem is also complicated.....
Ure problem depends on Re and AR (Archimede number).
All what i can tell u, is that for small Re<130, u need an axysimetric solver since u will get a vortex ring in front of the sphere.
For Re>130, the trajectory of ure bubble will no more be vertical: u need a 3D model. The trajectory form will depends on Ar number. straight line (but not in the vertica
For Re

[thecfduser]

sorry i pressed enter be4 i finished
For RE>130, ure trajectory will be a straight line (in any direction), a zig zag or an helocoidal...depending on Ar number.
For RE>270, it will be chaotic....

So, u can use axisymetric solver for small RE.
Else, u must use a 3D model
( u can use a 3D model with symetry plane if Re<170)

If not, u can not properly model the wake

[sega]

Which is exactly what he knows from (R. Clift at al ,1978) he mentioned he read.

[thecfduser]

heehehheeh
i think he is new in the domain
i think he red that for those sizes, the bubble stay spherical....

[Xilaf]

Yes. I'm really new in this domain. I have to study everything. Now I have atempt to Fluent, so I'm just trying to set the axisymetric solver.

[Xilaf]

I model only the situation, where Re is small and the trajectory will be straight line. So I expect, that the 2D sufficient for me.

[Xilaf]

Now I have a problem with the axisymetric solver.
The problem is in AMG solver: x-momentum, where is the divergence detected. Do you know, what's wrong?
Attached .jpg file.
Thanks

[Philipov]

Mesh quality, time step...could be anything.....

[Xilaf]

Yes. I've just run the calculation in axisymmetric space. The bubble rises 2x faster, than in planar space, but still slower, than the reality expects.