[emirust]

Dear Foamers,

I have a pretty easy case to set up, which I am basing on the damBreak tutorial, with the interFoam solver.

The goal is to simulate the fall of a liquid drop in another liquid.

Most aspects of the setup seems pretty straightforward to me, but I was wondering of the possibility of refining the mesh in a vertical column where the liquid sphere will fall. This is to limit the total number of cells of the mesh.

After looking a bit, it seems I should use the refineMesh utility. I am however a beginner in OpenFoam, and I don't know how to select the region to be refined. Based on the /multiphase/cavitatingFoam/ras/throttle case, I tried the following:

topoSetDict.1, to select a box where I wish to refine the region:

FoamFile
{
    version     2.0;
    format      ascii;
    class       dictionary;
    location    "system";
    object      topoSetDict.1;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

actions
(
    {
        name    c0;
        type    cellSet;
        action  new;
        source  boxToCell;
        sourceInfo
        {
            box (0.04 0.1 0) (0.06 0.25 0.01);
        }
    }
);

and the refineMeshDict file:

set             c0;

coordinateSystem global;

globalCoeffs
{
    tan1            ( 1 0 0 );
    tan2            ( 0 1 0 );
}

directions      ( tan1 tan2 );

useHexTopology  yes;

geometricCut    no;

writeMesh       no;

This will refine the whole mesh, multiplying the number of cells by 4 as it is a 2D case, but I want the refinement to be only in a limited region. Any help would be greatly welcomed!

[emirust]

Found out about it myself

For those wondering on the same thing, on need to run topoSet to extract the cells of interest (according to the coordinates of the box). (BTW, the name of the file should topoSetDict and not topoSetDict.1)

Then run refineMesh -dict -overwrite to specify you want to run the refineMesh utility linked to the refineMeshDict and also to overwrite the current mesh.

That's it!

[muhanad]

Hello,

I am curious to know, how you managed to create- at least a 2D droplet?

I am working in interFoam and was able to alter the geometry for rectangular cases, but still unable to create 2D drop!

Appreciate your help
regards

[muhanad]

Originally Posted by emirust

Found out about it myself

For those wondering on the same thing, on need to run topoSet to extract the cells of interest (according to the coordinates of the box). (BTW, the name of the file should topoSetDict and not topoSetDict.1)

Then run refineMesh -dict -overwrite to specify you want to run the refineMesh utility linked to the refineMeshDict and also to overwrite the current mesh.

That's it!

thanks for that, i ahve done it but since that time my simulation is running and still not ended up! it has taken up to the moment of writing the message more than 20 times of the original time
I am curious to see the result on paraView
by the way, what is the standard water droplet diameter?

Regards

[emirust]

Hello,

You may have a look at my master thesis: http://daim.idi.ntnu.no/masteroppgav...teroppgave.pdf

You can find most of your answers there, but let me know if there is something unclear that I haven't mentioned :-).

By the way, the thesis was a "failure", in the sense that the simulated case was not possible to simulate with the VOF method. Check the conclusion for some advice if you really want to simulate the same range.

Best regards,

Emile

[muhanad]

Thanks a lot, i will have a look at it.

I agree with you that VOF alone is insufficient, I came across a paper recently published where the simulation was possible with a coupled CLSVOF method, VOF with Level Set Method. VOF will not account properly for the diffusive flow physics within the interface, while continuity in VOF is maintained. Secondly, the level set method proofed to have sharp interface with a problem on conservation continuity equation, thus the solution is to build alpha in the VOF from phi in Level Set, and to build phi from Level set from alpha in VOF to guarantee both interface sharpness and continuity conservation..

Unfortunately, this paper has not elaborated on the numerical part to the extend that allowing some one to do the same.

Have you came across the above info before??