[RodriguezFatz]

Dear all,

I simulate some tank with air and water (mixture model) where the air rises and leaves at the top and water leaves the tank at the bottom.

When I test my grid with single phase water (and gravity), I get nice straight convergence. What I also tryed is to "play" with operating density in single phase and Fluent seems to have strong convergence issues, if the operating density is not the fluid density (which means that a buoyancy term appears in momentum equation as far as I understand it). However, using the correct density everything is fine, setting operating density to zero causes chaos: "turbulent viscosity limited to..." and divergence.

Now with mixture model activated, the simulation converges nicely as long as gravity is switched off. With gravity switched on, I get the same chaos as in the "single phase - zero operating density" case. That leads me to the assumption that the operating density is the crux of the matter. That is where I am stuck... everything I tryed does not work.

Can anyone help me?

[kbaker]

The operating density is significant in compressible flow however if you want to activate it it must be set for the lighter phase be aware not only to the operating density you need take care also for reference pressure location it must be set at the fluid which you set the operating density for it.

[RodriguezFatz]

I am sorry I did not mention it earlier, but my flow is incompressible. As I understand it, I do not need compressible air for buoyancy of air in water, right? How does your tip apply for incompressible flow? Also take air density as reference? Reference pressure should not matter, right?

[chaosh]

Originally Posted by RodriguezFatz

Dear all,

I simulate some tank with air and water (mixture model) where the air rises and leaves at the top and water leaves the tank at the bottom.

When I test my grid with single phase water (and gravity), I get nice straight convergence. What I also tryed is to "play" with operating density in single phase and Fluent seems to have strong convergence issues, if the operating density is not the fluid density (which means that a buoyancy term appears in momentum equation as far as I understand it). However, using the correct density everything is fine, setting operating density to zero causes chaos: "turbulent viscosity limited to..." and divergence.

Now with mixture model activated, the simulation converges nicely as long as gravity is switched off. With gravity switched on, I get the same chaos as in the "single phase - zero operating density" case. That leads me to the assumption that the operating density is the crux of the matter. That is where I am stuck... everything I tryed does not work.

Can anyone help me?

If gravity is a strong influence on your process, set operating density to "0". All this does is when you plot pressure you will see the hydrostatic pressure component in your simulation. The other things that might help your simulation are:

-When you start your simulation, keep your volume fraction under-relaxation low (about 0.5 or even lower) in your solution controls. As you simulation proceeds, gradually increase to 0.99. This I found is the biggest factor in my experience.

-Try using a coupled solver with low courant number (1), pressure and momentum under-relaxed to 0.25

-In the second phase properties panel, if you changed the phase diameter to something unreasonably large, this can happen as well. Move it back to default (its 10 microns i think).

Good luck

[ignacio]

I know that this is quite an old thread, but I have a similar issue and can't figure our how to fix it. I am using Fluent 17.0
My simulation is a water pool with a gas space above.

I am interested in computing the body forces in the liquid accurately. Therefore, to avoid round-off errors in the liquid buoyancy, my operating density should be the one of the liquid.
Using the gas density (as recommended by Fluent) leads to a completely unphysical results in the temperature distribution in the liquid.

The problem is that using the operating density of the liquid leads to a large pressure gradient in the gas space, which causes unphysically high velocities...

So, both options appear to be bad
1) Operating density = liquid --> good buoyancy in liquid, bad air convergence
2) Operating density = gas --> bad buoyancy in liquid, good air convergence

What should I do to have good buoyancy in the liquid and a stable gas space?

Thanks in advance,
Ignacio

[mohammad faizan]

Originally Posted by RodriguezFatz

Dear all,

I simulate some tank with air and water (mixture model) where the air rises and leaves at the top and water leaves the tank at the bottom.

When I test my grid with single phase water (and gravity), I get nice straight convergence. What I also tryed is to "play" with operating density in single phase and Fluent seems to have strong convergence issues, if the operating density is not the fluid density (which means that a buoyancy term appears in momentum equation as far as I understand it). However, using the correct density everything is fine, setting operating density to zero causes chaos: "turbulent viscosity limited to..." and divergence.

Now with mixture model activated, the simulation converges nicely as long as gravity is switched off. With gravity switched on, I get the same chaos as in the "single phase - zero operating density" case. That leads me to the assumption that the operating density is the crux of the matter. That is where I am stuck... everything I tryed does not work.

Can anyone help me?

Hello RodriguezFatz,
Have you got the result. I have the same problem, i am using mixture model for nano fluid similation inside pipe. Could you please help me if you got your result?
Thanks in advance