[Hr_kules]

Hey all,


i have a geometry of a plate with an angular step and and surrounding fluid volume. The stls were created in python because i exporting them from Ansys led to certain points not matching each other. The stls themself are attached. The result of running snappyHexMesh are also attached. At first they look somewhat okay, but i get 26 wrong faces due to orthogonality.
Also the boundaries are a complete mess after spliting the mesh:
boundaries fluid:

/*--------------------------------*- C++ -*----------------------------------*\
  =========                 |
  \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox
   \\    /   O peration     | Website:  https://openfoam.org
    \\  /    A nd           | Version:  9
     \\/     M anipulation  |
\*---------------------------------------------------------------------------*/
FoamFile
{
    format      ascii;
    class       polyBoundaryMesh;
    location    "constant/fluidcold/polyMesh";
    object      boundary;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

3
(
    wallfluidcold
    {
        type            wall;
        inGroups        List<word> 1(wall);
        nFaces          108;
        startFace       3581709;
    }
    wallsolid
    {
        type            wall;
        inGroups        List<word> 1(wall);
        nFaces          199224;
        startFace       3581817;
    }
    fluidcold_to_solid
    {
        type            mappedWall;
        inGroups        List<word> 
3
(
wall
mappedPatch
baffles1
)
;
        nFaces          195024;
        startFace       3781041;
        sampleMode      nearestPatchFace;
        sampleRegion    solid;
        samplePatch     solid_to_fluidcold;
    }
)

// ************************************************************************* //

boundaries solid

/*--------------------------------*- C++ -*----------------------------------*\
  =========                 |
  \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox
   \\    /   O peration     | Website:  https://openfoam.org
    \\  /    A nd           | Version:  9
     \\/     M anipulation  |
\*---------------------------------------------------------------------------*/
FoamFile
{
    format      ascii;
    class       polyBoundaryMesh;
    location    "constant/solid/polyMesh";
    object      boundary;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

4
(
    inletcold
    {
        type            patch;
        nFaces          80283;
        startFace       6892163;
    }
    outletcold
    {
        type            patch;
        nFaces          80284;
        startFace       6972446;
    }
    wallfluidcold
    {
        type            wall;
        inGroups        List<word> 1(wall);
        nFaces          217620;
        startFace       7052730;
    }
    solid_to_fluidcold
    {
        type            mappedWall;
        inGroups        List<word> 
3
(
wall
mappedPatch
baffles1
)
;
        nFaces          195024;
        startFace       7270350;
        sampleMode      nearestPatchFace;
        sampleRegion    fluidcold;
        samplePatch     fluidcold_to_solid;
    }
)

// ************************************************************************* //

As you can see the boundaries don't match up at all. This is frustrating as i plan to automatize the stl creation and meshing as well as the calculation and this case, was acutally just a tester. The allrun script looks like this:

#!/bin/sh
cd "${0%/*}" || exit 1    # run from this directory

#Color for messages
GREEN='\e[0;32m'
NC='\e[0m'

echo "${GREEN}create background mesh$NC"
blockMesh > meshLog
echo "${GREEN}create edge mesh$NC"
surfaceFeatures >> meshLog
echo "${GREEN}run snappyHexMesh$NC"
snappyHexMesh -overwrite >> meshLog
echo "${GREEN}create baffles for interface$NC"
createBaffles -overwrite >> meshLog
echo "${GREEN}split mesh along interface$NC"
splitMeshRegions -cellZones -defaultRegionName solid -overwrite >> meshLog
echo "${GREEN}create singles files for paraFoam$NC"
paraFoam -touchAll

Any help is really appreciated and very well needed, as i tried everything i could think of, but i am still stuck.



Thank you!

[Antimony]

Hi,

Can you explain what you mean by

As you can see the boundaries don't match up at all

?

If you see the interface, they seem to have the same number of faces. So in that sense they match up.

Also, you mention that the mesh quality is not good? Can you post the output of checkMesh as well as the snappyHexMeshDict that you used?

Hope this helps.

Cheers,
Antimony

[Hr_kules]

Hey,


yes i can explain; so for the region fluidcold i want following boundaries as specified in my stl:
-inletcold
-outletcold
-wallfluidcold
-fluidcold_to_solid
and for the solid region:
-wallsolid
-solid_to_fluidcold
in the code snippets the boundaries appear more then once for each region as well as that in and outlet are somehow in the solid region.


I have attached the snappyhexmeshdict and the log of all mesh operations. The checkMesh run is the last entry in the log.


Thank you!

[Hr_kules]

Hi Antimony,


have you had any chance to llok in the uploaded files? I would really appreciate some guidance considering my problems.


Thank you and have a nice day,


Nico

[Antimony]

Hi,

I took a quick look. Seems like you are getting 2 regions which is inline with what you want. And the mesh quality is fine. The 26 faces are marginally beyond the default value of 65 deg, so that really should not be causing any issues.

Unfortunately, at the moment it isn't clear where the issue is. Given that you have multiple STLs + baffles, it is difficult to pinpoint without being able to see the case...

Have you looked at the conjugate heat transfer tutorial cases as a guide to setting up multiple region simulations?

Cheers,
Antimony

[Hr_kules]

Hey,


if it helps, i have attached to full case. My main concern is the assignment of the boundaries after spliting the mesh.


I have looked up the tutorials before starting this case, especially the shell and tube heat exchanger example. I am somewhat confused why the assignment of the boundaries works there considering the geometry of a shell and tube heat exchanger is quite a bit more complex then two blocks with a chevron step on it.



Thank you and have a nice day!

[Hr_kules]

Hi,
i know this is a bit late, but i could pin point the problem considering the multiregion meshing. It seems like the cellZones are not assigned to right region. However i have no idea how to fix this..
Maybe you see this and can enlighten me?

[vittoria]

Originally Posted by Hr_kules

Hey,


yes i can explain; so for the region fluidcold i want following boundaries as specified in my stl:
-inletcold
-outletcold
-wallfluidcold
-fluidcold_to_solid
and for the solid region:
-wallsolid
-solid_to_fluidcold
in the code snippets the boundaries appear more then once for each region as well as that in and outlet are somehow in the solid region.


I have attached the snappyhexmeshdict and the log of all mesh operations. The checkMesh run is the last entry in the log.


Thank you!

I'm currently facing the same problem with the boundaries, were you able to fix it, and if yes how did you do it?


Thanks you

[Hr_kules]

Hi, i was unfortunately never able to solve the problem. My conclusion was to not use snappyHexMesh for multi region meshes ever again. I changed to Salome for more complex meshes; easy meshes i have done with GMSH.