[celestial]

I create a box in Gmsh under Modules/Geometry. This is the .geo file:

// Gmsh project created on Wed Nov 13 14:12:11 2019
//+
Point(1) = {-0.3, -0.7, 0, 1.0};
//+
Point(2) = {0.2, -0.7, 0, 1.0};
//+
Point(3) = {0.2, -0.5, 0, 1.0};
//+
Point(4) = {-0.1, -0.5, 0, 1.0};
//+
Line(1) = {3, 4};
//+
Line(2) = {1, 4};
//+
Line(3) = {1, 2};
//+
Line(4) = {2, 3};
//+
Curve Loop(5) = {4, 1, -2, 3};
//+
Plane Surface(1) = {5};
//+
Extrude {0, 0, 1} {
Surface{1};
}
//+
Physical Surface("front") = {26};
//+
Physical Surface("back") = {1};
//+
Physical Surface("in") = {13};
//+
Physical Surface("out") = {21};
//+
Physical Surface("top") = {17};
//+
Physical Surface("bot") = {25};
//+
Physical Volume("vol") = {1};




In Gmsh, I click on 3D under Modules/Mesh and then export the .msh file as ascii version 2 (gmshToFoam cannot read version 4) checking Save all elements ( but not Save parametric coordinates since gmshToFoam complains about that)
It resides in the case directory of openfoam. Here is its contents:

$MeshFormat
2.2 0 8
$EndMeshFormat
$PhysicalNames
7
2 1 "front"
2 2 "back"
2 3 "in"
2 4 "out"
2 5 "top"
2 6 "bot"
3 7 "vol"
$EndPhysicalNames
$Nodes
9
1 -0.3 -0.7 0
2 0.2 -0.7 0
3 0.2 -0.5 0
4 -0.1 -0.5 0
5 0.2 -0.7 1
6 0.2 -0.5 1
7 -0.1 -0.5 1
8 -0.3 -0.7 1
9 -0.02261904761904762 -0.6166666666666666 0.5357142857142857
$EndNodes
$Elements
43
1 15 2 0 1 1
2 15 2 0 2 2
3 15 2 0 3 3
4 15 2 0 4 4
5 15 2 0 5 5
6 15 2 0 6 6
7 15 2 0 10 7
8 15 2 0 14 8
9 1 2 0 1 3 4
10 1 2 0 2 1 4
11 1 2 0 3 1 2
12 1 2 0 4 2 3
13 1 2 0 7 5 6
14 1 2 0 8 6 7
15 1 2 0 9 7 8
16 1 2 0 10 8 5
17 1 2 0 12 2 5
18 1 2 0 13 3 6
19 1 2 0 17 4 7
20 1 2 0 21 1 8
21 2 2 0 1 1 2 4
22 2 2 0 1 2 3 4
23 2 2 0 14 2 3 6
24 2 2 0 14 5 2 6
25 2 2 0 18 3 4 6
26 2 2 0 18 6 4 7
27 2 2 0 22 4 1 7
28 2 2 0 22 7 1 8
29 2 2 0 26 1 2 8
30 2 2 0 26 8 2 5
31 2 2 0 27 5 6 7
32 2 2 0 27 8 5 7
33 4 2 0 1 2 6 3 4
34 4 2 0 1 4 1 2 9
35 4 2 0 1 8 7 5 9
36 4 2 0 1 8 2 1 9
37 4 2 0 1 5 2 8 9
38 4 2 0 1 7 1 4 9
39 4 2 0 1 8 1 7 9
40 4 2 0 1 7 6 9 4
41 4 2 0 1 7 9 6 5
42 4 2 0 1 2 9 6 4
43 4 2 0 1 2 6 9 5
$EndElements




Finally, after running gsmhToFoam, the boundary file resides in folder constant/polyMesh :




/*--------------------------------*- C++ -*----------------------------------*\
| ========= | |
| \\ / F ield | foam-extend: Open Source CFD |
| \\ / O peration | Version: 4.0 |
| \\ / A nd | Web: http://www.foam-extend.org |
| \\/ M anipulation | |
\*---------------------------------------------------------------------------*/
FoamFile
{
version 2.0;
format ascii;
class polyBoundaryMesh;
location "constant/polyMesh";
object boundary;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

1
(
patch0
{
type patch;
nFaces 12;
startFace 16;
}
)


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



As you can see it creates a patch name patch0. So, totally disregarding the physical surfaces !

Can anybody see what is going on ? I am very curious about this.

Thanks, Marc

[celestial]

Well, forget about Gmesh 4.3.0 and saving the mesh in version 2 format: I ran Gmesh 12.2.0 and now the physical surfaces were correctly transferred to patches for that particular convex shape. Unfortunately, only 2
out of 3 surfaces were transferred for another shape with a tiny dent: the 2 surfaces comprising the dent were successfully transferred but not the surface for the shape surface minus the dent surfaces. And for a multiply connected shape ( i.e. having a hole inside) the outcome is a lot worse: gmshToFoam complains that the mesh has no cells (volumes) even though after clicking Mesh/3D in Gmesh but before saving the mesh, the bottom command window would clearly state that there are a positive number of volumes in the mesh ! So at this point, I don't know which Gmesh version since 2002 (if any) fits the bill.

[celestial]

Well, I cried wolf too fast for the shape with a tiny dent: it was simply a typo

Physical Surface("airwing") += {3, 2, 1, 5, 4};

The += should have been just =


And sorry I meant Gmsh 2.12.0 all along not 12.2.0 obviously

Now I will try to tackle the more serious problem that remains and I will keep you posted. Thanks for viewing

[celestial]

How careless and sloppy one can be.

It was again a typo with the + sign:

Physical Volume("vol") += {1};

I meant of course

Physical Volume("vol") = {1};

So there is nothing special about multiply connected regions:

Now gmshToFoam will yield the right constant/polyMesh files

i.e. I now have a 3D mesh and all physical surfaces were correctly transferred to the corresponding patch names in the boundary files


My most sincere apologies for wasting everybody's time viewing this post.