I'm using a commercial CFD software and I need to model a liquid film flowing over a sphere using VOF. Air is flowing in the same direction at a higher velocity ( not too high).

I want to change the drag forces in the gas-liquid and liquid-air interfaces. I suspect that the equations I want to implement must hold only in the gas-solid interface and in the free surface ( volume fraction<>0 or 1).

I asked the technical staff on how to do it, as only answer they wondered why should I want to change the drag forces in VOF.

I want to change them, but it rised questions on if what I'm doing is right.

Can somebody help solve my doubts?

Thanks a lot in advance.

some1

I did a multi-phase problem on Fluent. And I agree with "the technical staff" on why you have to or not change the drag forces in VOF. Also Vol Fraction is between 0 and 1. When it comes to computer simulations(i.e. CFD) start with a simple and quick solution approach, (use the default values whenever possible) and when you get your initial solution, you can then play around with your approach to get the solution that you want. This is the only way(I hope I'm wrong...) you're going to find the answer to your "Doubt on VOF". Can you send me a picture of your problem? Thanks. I hope I was of some help and Go with the Flow, Edward

Thanks a lot for your answer. I have already solved the problem with the default values, I'm in the next step as you mention.

I'm sending you the figure.

Thanks again

some1

Arturo; Thanks for the picture. Now, I know what you're up to. I agree with your initial assessment of your initial solution. Here are some more hints:

It is possible for you to animate your solution. For example, for every 10 iteration you make, save a hardcopy of the window containing the resulting plot as a tiff file. (Make sure that every window is the same size) Then use a graphics package to assemble all the tiff files into 1 animated file. You can only do this manually. You can't write this into a script and run Fluent in background mode. It's nice to have an animated solution to an unsteady problem. You'll really see what's happening.

With regard to "the drag forces in the liquid-solid and gas-liquid interfaces..." You may have to change your mesh so that you have more cells in the region close to liq-solid & gas-liquid interfaces to get a better answer for everything including the drag forces. Try using TGrid to do this, Gambit will not do this. A really easy and dirty way is to go back to Gambit and increase the density of your mesh. But this method slows Fluent and if you get carried away, your Admin will not like it if he finds out that you're using almost all the resources available.

Just keep on going with Fluent, it's a really nice package. The only thing I hate about it is its mesh generators(take a look at any CFD book or paper, you'll know what I mean). If you have any more questions, just let me know. Go with the Flow, Edward

You can't change the drag of the VOF free surface model. The drag is infinite because VOF is basically the homogenous limit of a full multiphse model. To change the drag you really need to switch to a full multiphase model so that there is a drag term for you to modify.

Neale

thanks!!!! I will try your hints!!

some1

Thanks Neale That was my doubt Do you have any advice?

I really appreciate your input.

some1

Dear Neale your answer made me think about some possible strategies. Do you think is possible to use the predicted liquid surfaces to perform further simulations with other model?

thnx

some1

1) There is a major problem concerning drag between liquid and gas using the VOF method (implemented in Fluent, STAR-CD, CFX and others). 2) The drag is a result of the simulation itself and can only be as accurate as the liquid/gas interface is resolved. 3) The interface unfortunately smears over 3 to 6 cells so you can more or less forget drag results 4) This is not the case in a different method called MAC (Marcer And Cell). But there is no commercial code with that method implemented

Thanks Helge, it seems a weekness of VOF, from what I've read the model solves a single momentum equation and it can be added some forces, i.e. surface tension.

For me it would be natural to add drag forces but from the formulation is not evident.

new1

Helge,

The drag is infinte in a free surface VOF model, i.e. there is no slip velocity between the phases. So, it's not that there is a problem with the commercial codes, it's a fundamental limitation of the model, no matter who implements it.

If drag is important, then in CFX-4 and CFX-5.5 you can use the full multiphase model instead, which has a slip velocity drag model. This works fine.

Neale.

Single phase momentum equation and a continuity equation accommodating the density (weighted average) are the basis of VOF - am I correct in my understanding ? Suppose I am modeling the surface disturbance in a liquid (say water) exposed to air, using VOF. How do I take care of the viscosity ? Is it the weighted average too ? Is that a right proposition? Viscosity is intensive property - can I do that? Has there been any attempt to evaluate/estimate the prospective numerical diffusion ? Else, what are we predicting ? At this stage, I am not talking of Newtonian behavior or not. newtonian

To your first point. I doubt that the drag is infinite in a VOF formualtion. Of course the relative velocity between the two phases is zero. But that is also the case between a fluid and a wall in single phase flow. Nevertheless there is drag between the fluid and the wall. So it should be possible to calculate a drag between the two fluids.

To your second point. If you want to get the drag out of a simulation you should not use a velocity drag model in a full multiphase model because you are using an emperical formula for the drag i.e. you are using the result as input.