[er_ijaz]

Hi all,

I'm modelling urban airflow using buoyantsimplefoam. We are using actual terrain instead of flat terrain. I created mesh using the following command

decomposePar

mpirun -np 64 snappyHexMesh -parallel -overwrite

reconstructParMesh -constant

checkMesh

Log snappyhexmesh ends with

Writing mesh to time constant
Wrote mesh in = 350.41 s.
Mesh snapped in = 2558.65 s.
Checking final mesh ...
Checking faces in error :
non-orthogonality > 60 degrees : 0
faces with face pyramid volume < 1e-13 : 0
faces with face-decomposition tet quality < 1e-15 : 0
faces with concavity > 80 degrees : 0
faces with skewness > 4 (internal) or 20 (boundary) : 0
faces with interpolation weights (0..1) < 0.05 : 0
faces with volume ratio of neighbour cells < 0.01 : 0
faces with face twist < 0.02 : 0
faces on cells with determinant < 0.001 : 0
Finished meshing without any errors
Finished meshing in = 5887.69 s.
End

Finalising parallel run
After reconstructParMesh and while checking mesh

I get one quality failed

Checking geometry...
Overall domain bounding box (0 0 -5.739853e-14) (5500 5500 1000)
Mesh has 3 geometric (non-empty/wedge) directions (1 1 1)
Mesh has 3 solution (non-empty) directions (1 1 1)
Boundary openness (1.53412e-16 -3.11233e-15 1.939486e-13) OK.
Max cell openness = 4.438246e-16 OK.
Max aspect ratio = 9.576651 OK.
Minimum face area = 0.007753621. Maximum face area = 421.7486. Face area magnitudes OK.
Min volume = 0.05158672. Max volume = 6093.245. Total volume = 3.007628e+10. Cell volumes OK.
Mesh non-orthogonality Max: 60.00908 average: 3.920135
Non-orthogonality check OK.
Face pyramids OK.
***Max skewness = 7.219928, 10 highly skew faces detected which may impair the quality of the results
<<Writing 10 skew faces to set skewFaces
Coupled point location match (average 0) OK.

Failed 1 mesh checks.
I'm not sure why mesh fails after reconstructing

When I run the simulation with single core the simulation starts without error

But while decompose again and run parallel my decomposition fails with errors like

--> FOAM FATAL IO ERROR:
error in IOstream "/home/users/nus/ijaz01/PinnacleCaseStudy1_buoyantsimplefoam/20230920/case2_reducedmesh/simulation1/processor16/constant/polyMesh/faces" for operation Foam::Ostream& Foam:perator<<(Foam::Ostream&, int64_t)

file: /home/users/nus/ijaz01/PinnacleCaseStudy1_buoyantsimplefoam/20230920/case2_reducedmesh/simulation1/processor16/constant/polyMesh/faces at line 2867957.

or

[2] --> FOAM FATAL ERROR:
[2] Cannot find file "points" in directory "polyMesh" in times "0" down to constant

my decomposePar dict is as below

numberOfSubdomains 8;

method scotch;

simpleCoeffs
{
n ( 2 2 2 );
delta 0.001;
}

hierarchicalCoeffs
{
n ( 1 1 1 );
delta 0.001;
order xyz;
}

manualCoeffs
{
dataFile "";
}

distributed no;

roots ( );

I'm using openfoam 1912, is there any issue with decomposition and reconstructPar that causes error while trying to run the simulation in parallel?

[Yann]

Hello,

If you want to compare your mesh quality before/after reconstructing your mesh, you should rather run checkMesh twice. Something like:

decomposePar

mpirun -np 64 snappyHexMesh -parallel -overwrite

mpirun -np 64 checkMesh -parallel

reconstructParMesh -constant

checkMesh

Please note snappyHexMesh uses meshQualityControls parameters defined in snappyHexMeshDict while checkMesh uses a default set of quality criteria. You can use checkMesh with the -meshQuality option to provide your own set or quality criteria rather than using the defaults ones.

10 faces with a skewness about 7 is usually not a big deal for the simulation. (On complex geometry it might not always be possible to totally get rid of skewed faces).

[2] --> FOAM FATAL ERROR:
[2] Cannot find file "points" in directory "polyMesh" in times "0" down to constant

This error is usually related to a mesh size mismatch (for instance, data is written in 0 directory, but the size does not match the size of the mesh in constant)

Something seems to be off about the information your provide: you run snappy on 64 cores, but your decomposeParDict mentions 8 subdomains only .

--> FOAM FATAL IO ERROR:
error in IOstream "/home/users/nus/ijaz01/PinnacleCaseStudy1_buoyantsimplefoam/20230920/case2_reducedmesh/simulation1/processor16/constant/polyMesh/faces" for operation Foam::Ostream& Foam:perator<<(Foam::Ostream&, int64_t)

file: /home/users/nus/ijaz01/PinnacleCaseStudy1_buoyantsimplefoam/20230920/case2_reducedmesh/simulation1/processor16/constant/polyMesh/faces at line 2867957.

This error mentions processor16, suggesting there are indeed more than 8 subdomains.

And one last thing: reconstructing the mesh is optional. You can run a whole case and post-process it without having to reconstruct anything.

Hopes this helps,
Yann