CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Filling a cavity (

Nuno September 22, 2008 08:40

Filling a cavity

I am trying to modelling a filling of a cavity. This cavity is full of air, and i will try to fill with water. I used a fluid pair air/water with the volume fraction at inlet of 1 to water and 0 to air. I try but the results are that the volume fraction remains constant, so dont change with time. This is a transient analyse.

Some one can help me?


Rui September 22, 2008 10:26

Re: Filling a cavity

What other conditions are you imposing at the inlet? A given inlet velocity?

Do you have an outlet for the air to escape?

Nuno September 22, 2008 11:50

Re: Filling a cavity
yes i have an outlet for the air to escape. I put like a pressure 0. At inlet i put a velocity.

Any ideia?

I saw a tutorial to simulate the smoke but they insert a new variable to simulate the concentration of smoke and they only work with one fluid, and they use a transportation equation. CFX have this transportation equation pre-define for other pre-difine variables?

Rui September 22, 2008 15:33

Re: Filling a cavity

What are your initial conditions? Are the initial volume fractions of water 0 and of air 1 (full of air)? If they are, and if at the inlet you have volume fraction of water 1, and you have an outlet, there should be water flowing into the domain.

But what do you mean by the volume fractions don't change with time? Do they remain the same as what you imposed for the initial conditions? Or do you mean they change at the beginning, but after a while they become constant?

Is your simulation converging? This is just a guess, as I don't know the settings of your simulation, but it might be that your simulation isn't converging (or even it didn't converge in only one of the timesteps) and you have huge velocities inside your domain ... and what you're getting from the simulation is just rubbish. Make a plot of velocities vectors, and see if they are what you'd expect them to be. But you should always make sure that the simulation is converging.

Apart from being transient, your simulation has nothing to do with that Smoke tutorial. Actually, it shall be similar to the "Free Surface Flow Over a Bump" tutorial: Two Phases (Water and Air), Free Surface, Buoyancy

If you look at the documentation, you'll see that, for Multiphase Flows (Solver Theory - Multiphase Flow Theory), CFX solves a continuity equation for each phase, which may be regarded as transport equations for the volume fractions.


Nuno September 23, 2008 06:57

Re: Filling a cavity

I made a several mistakes and wrong explanation.

My initial conditions are are the initial volume fractions of water 0 and of air 1 (full of air) and at inlet a velocity for water and volume fraction of 1.

There is water flowing. But i cant see the volume fraction changing in time. What i mean. I want to see the water filling the geometry, so, this mean, i want to see the the volume fraction changing in time.

I refer the smoke tutorial because they use a additional variable and add a transport equation, and in those tutorial we can see the concetration of the smoke changing in identical to what i want but with two fluids...and a free surface.

Thank for your help

Rui September 23, 2008 10:53

Re: Filling a cavity

If there's water flowing in, and no (or less) water flowing out, then the volume of water inside the domain has to increase, right?

So, what do you actually see? Can you post here a couple of pictures? (you can upload them to any web site, for example: - usually I use -, and post here the links)

rohit September 23, 2008 11:08

Re: Filling a cavity
Dear Nuno,

This may sound stupid, but i have once experienced something similar, you just try to initialize the domain with 0.001%water and 0.999% air rather tha 0 and 1. i hope this time you will see the cavity will be filling with water

Nuno September 24, 2008 10:18

Re: Filling a cavity

I get some results, doing like rohit said. But there is another thing strange...only the water change in space and time. The air remains constant and that is very strange. I need to acoplate the both volume fraction?

Sorry for my stupid questions


Nuno September 25, 2008 09:16

Re: Filling a cavity

i will put some pic of the results.

This is from water

This is from air

I dont have a very strong discretization of the problem. My aim is to try to simulate and get some experience to do the filling.

The initial volume fractions of water 0,0001 and of air 0,9999 (full of air). At the inlet you have volume fraction of water 1, and you have an outlet. The results seems that both volume fraction arent coupled. I put this like a standard free surface model.

Is very strange because sometimes seems that the Volume fraction at domain are right, because i put 0 to the water and in results they are fill of water. For an steady analysis.

Rui September 26, 2008 12:07

Re: Filling a cavity

I don't know what's the problem. The volume fraction of Air + the volume fraction of Water should be 1. In you pictures they are 0.9999 (Air) and between 0.35 and 0.49 (Water)! Are you sure the variables values are from the same instant? If you're getting the results from a partial (minimum) transient file, and if you didn't select the Air volume fraction, for example, when you defined the transient results in Pre, than CFX-Post shows the Air volume fraction from the closest instant it is available.

And is your simulation converging?

Nuno October 1, 2008 10:52

Re: Filling a cavity

I could get some results. My problem is that i was using a results preview results file from a steady analisys, so the problem was calculated based on that result. The problem is that those results was a full geometry fill.

Thank guys!!!

Rui October 1, 2008 19:23

Re: Filling a cavity

All times are GMT -4. The time now is 17:35.