[voingiappone]

Hello everybody,

I'm forced to ask here a very common question... already being asked many times, but I cannot realize how to get started in the right way. Please be patient.....

Situation:
I want to run a simulation in a bubble column (cylinder) with a somewhat complicated geometry. To make things easy, I started running the tutorial provided and everything goes (of course) fine. Browsing the files I realized that the p, alpha, Ua and Ub profiles are already given in the definition folder.

Problem:
In my case the fluid will be steady when I will start bubbling air in, so this definition of the variables simply doesn't work. Moreove, how can I fugure out a distribution of U,p and alpha before actually running the calculations?!?!

What I did was to purge from the files the data from the examples and correct the parameters for having a steady fluid, and a 0.1 m/s gas inlet. My config files are these.

Alpha

dimensions [0 0 0 0 0 0 0];

internalField uniform 0;

boundaryField
{
Inlet
{
type fixedValue;
value uniform 0.5;
}

Outlet
{
type fixedValue;
value uniform 0;
}

Walls
{
type zeroGradient;
}

}

pressure

dimensions [1 -1 -2 0 0 0 0];

internalField uniform 0;

boundaryField
{
Inlet
{
type zeroGradient;
}

Outlet
{
type fixedValue;
value uniform 0;
}

Walls
{
type zeroGradient;
}


}

Ua

dimensions [0 1 -1 0 0 0 0];

internalField uniform (0 0 0);

boundaryField
{
Inlet
{
type fixedValue;
value uniform (0 0 0.1);
}

Outlet
{
type zeroGradient;
}

Walls
{
type fixedValue;
value uniform (0 0 0);
}


}

Ub

dimensions [0 1 -1 0 0 0 0];

internalField uniform (0 0 0);

boundaryField
{
Inlet
{
type zeroGradient;
}

Outlet
{
type zeroGradient;
}

Walls
{
type fixedValue;
value uniform (0 0 0);
}


}

I'm making the fluid move along the z axis (±not like in the example, along y).

I don't know what's going on in the simulation but as I look at the results in the paraview, I see that alpha is already non-zero after 1 second almost everywhere in the column, Ua and Ub start rising in value from the top and other strange behaviours like this.

I also tried to reply wat was written in this forum post but without any luck...

I don't know what I am getting wrong even if I strongly suppose that it resides in these configuration files. If somebody could help me getting out from this....

Thank you in advance for your help

Luca

[TomA]

Hi Luca,

since you initialise your entire alpha field with 0. ("internalField uniform 0;") there is no liquid in your bubble column. You can use funkySetFields to change the alpha value in parts of your domain.Hope this helps.

Tom

[voingiappone]

TomA, thanx for your reply.
I believe that this you point out is one of more complicated problems affecting the boundary conditions....
However, I have looked through the documentation about bubbleFoam located here and it clearly states that alpha is the "dispersed pjase fraction". I interpreted this as being the gas fraction... which means

alpha=Vg/(Vg+Vl)

Hence, alpha=1 ---> 100% gas and alpha=0 ---> 100% liquid
Maybe I'm missing something though.....

However, I realized yesterday how to bound with a box an entire region ad how to change the value for alpha in that region, so I created a column with a top part with a=0 and a top one with a=1....
This case fails in 0.16 secs... with an infinite Courant number... Think I have to look for more adequate boudary conditions.

Anybody wants to share some tips???

Thank you

[voingiappone]

I have to admit that I gave up.... I moved to interFoam.
However, I have questions about that solver too.... I posted a new thread here.
If a solution comes, pleas write it anyway. It might be useful.

[Aurelien Thinat]

Hi voingiappone,

Where did you put the alpha = 1 ? At the bottom part of the column ?
"Alpha = 1" means there is no continuous phase right there. You may use a VOF solver for this kind of problems. BubbleFoam works for only 1 continuous phase and only 1 dispersed phase.

[voingiappone]

Thank you for your comment Aurelien Thinat.

The alpha file shown at the top of the thread is the one I used on my first trials. As you can see it's a 100% liquid filled column with an inlet on the bottom where alpha is set to be 0.5.
After that I realized that the boundary problem could be related to the liquid in the top part "flowing out" and learned how to make the column half liquid and half gas with the setFields utility. This changed the things a little but to no avail. Of course the liquid was on the bottom and the gas on the top .


As you suggest I already moved to the VOF approach (interFoam) and I solved my problems, at least regarding the boundary conditions and the ab-initio behavior.
To be sincere, I'm solving this problem because I need to measure the fluid speed on the top of the reactor so I wanted something less cpu intensive than a VOF as I'm not interested in the bubble shape, behavior, etc. I realized that it is not possible with my limited knowledge of the eulerian solver..... and so I had to move.

[miles_davis]

Hi,

I am trying to simulate air injection with a nozlle in a watter box.
I am struggling with twophaseeulerfoam.
It seems I am no the only one.
My question is what did you do finally?
voingiappone, did you finally use interfoam or interdyfoam? was it easy to handle? or did you finally manage to use bubblefoam?

Thanks for your help