[saifee]

Following is a checkMesh report of a mesh I've started to work with. I meshed it using Gmsh. I know it finally says mesh is OK, but it does have some non-Ortho faces. And I've had simulations in the past going unstable even though the mesh was OK. And since this is a vey large mesh, I'd like to have some assurance before I start the simulation (pimpleFoam).

1. Is this mesh good?
2. What are the thresholds for different parameters to be considered as a good mesh?
3. Are hexahedra cells better than tetrahedra cells?

/*---------------------------------------------------------------------------*\
  =========                 |
  \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox
   \\    /   O peration     | Website:  https://openfoam.org
    \\  /    A nd           | Version:  10
     \\/     M anipulation  |
\*---------------------------------------------------------------------------*/
Build  : 10

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create polyMesh for time = 0

Time = 0s

Mesh stats
    points:           640148
    faces:            6629172
    internal faces:   6222772
    cells:            3212986
    faces per cell:   4
    boundary patches: 4
    point zones:      0
    face zones:       0
    cell zones:       1

Overall number of cells of each type:
    hexahedra:     0
    prisms:        0
    wedges:        0
    pyramids:      0
    tet wedges:    0
    tetrahedra:    3212986
    polyhedra:     0

Checking topology...
    Boundary definition OK.
    Cell to face addressing OK.
    Point usage OK.
    Upper triangular ordering OK.
    Face vertices OK.
    Number of regions: 1 (OK).

Checking patch topology for multiply connected surfaces...
    Patch               Faces    Points   Surface topology                  
    a               	  400576   200447   ok (non-closed singly connected)  
    b         	          656      363      ok (non-closed singly connected)  
    c            	  2576     1351     ok (non-closed singly connected)  
    d            	  2592     1361     ok (non-closed singly connected)  

Checking geometry...
    Overall domain bounding box (-0.019484 -0.081107 -0.045986) (0.016059 0.017675 0.027626)
    Mesh has 3 geometric (non-empty/wedge) directions (1 1 1)
    Mesh has 3 solution (non-empty) directions (1 1 1)
    Boundary openness (2.25262e-17 -2.91212e-17 -2.57751e-16) OK.
    Max cell openness = 2.96517e-16 OK.
    Max aspect ratio = 25.1844 OK.
    Minimum face area = 2.20319e-09. Maximum face area = 4.69207e-07.  Face area magnitudes OK.
    Min volume = 8.0789e-14. Max volume = 9.1916e-11.  Total volume = 3.20018e-05.  Cell volumes OK.
    Mesh non-orthogonality Max: 86.6204 average: 15.0856
   *Number of severely non-orthogonal (> 70 degrees) faces: 451.
    Non-orthogonality check OK.
  <<Writing 451 non-orthogonal faces to set nonOrthoFaces
    Face pyramids OK.
    Max skewness = 1.38541 OK.
    Coupled point location match (average 0) OK.

Mesh OK.

End

[AtoHM]

Not all checkMesh thresholds are hard limits, but the nonOrthogonality is an exception. You should never go above the stated 70 degrees, even above 65 can be critical sometimes. So no - with this mesh I would not expect accurate results.

[saifee]

Originally Posted by AtoHM

Not all checkMesh thresholds are hard limits, but the nonOrthogonality is an exception. You should never go above the stated 70 degrees, even above 65 can be critical sometimes. So no - with this mesh I would not expect accurate results.

I see. So does the number of faces >70 degrees matter? Like a certain percentage of the total number of faces? Or even if 1 face has this non-orthogonality issue, it's bad enough?

[saifee]

Originally Posted by AtoHM

Not all checkMesh thresholds are hard limits, but the nonOrthogonality is an exception. You should never go above the stated 70 degrees, even above 65 can be critical sometimes. So no - with this mesh I would not expect accurate results.

Also, in such cases, does the nNonOrthogonalCorrectors do anything to help?

[AtoHM]

Yes, one such cell can make your simulation invalid. You should analyse where this occurs in your mesh and try to work against it using refinements, alter slightly or simplify the geometry etc.


https://www.simscale.com/docs/simula...al-correctors/


However, I'd rather not fix my simulation up with correctors, but improve my mesh.

[saifee]

I tried making some improvements to the mesh and this is the checkMesh results:

/*---------------------------------------------------------------------------*\
========= |
\\ / F ield | OpenFOAM: The Open Source CFD Toolbox
\\ / O peration | Website: https://openfoam.org
\\ / A nd | Version: 10
\\/ M anipulation |
\*---------------------------------------------------------------------------*/
Build : 10
Exec : checkMesh
Date :
Time :
Host :
PID : 1930324
I/O : uncollated
Case :
nProcs : 1
sigFpe : Enabling floating point exception trapping (FOAM_SIGFPE).
fileModificationChecking : Monitoring run-time modified files using timeStampMaster (fileModificationSkew 10)
allowSystemOperations : Allowing user-supplied system call operations

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create polyMesh for time = 0

Time = 0s

Mesh stats
points: 97774
faces: 908360
internal faces: 807408
cells: 428942
faces per cell: 4
boundary patches: 4
point zones: 0
face zones: 0
cell zones: 1

Overall number of cells of each type:
hexahedra: 0
prisms: 0
wedges: 0
pyramids: 0
tet wedges: 0
tetrahedra: 428942
polyhedra: 0

Checking topology...
Boundary definition OK.
Cell to face addressing OK.
Point usage OK.
Upper triangular ordering OK.
Face vertices OK.
Number of regions: 1 (OK).

Checking patch topology for multiply connected surfaces...
Patch Faces Points Surface topology
a 99552 49864 ok (non-closed singly connected)
b 296 169 ok (non-closed singly connected)
c 604 342 ok (non-closed singly connected)
d 500 282 ok (non-closed singly connected)

Checking geometry...
Overall domain bounding box (-0.019613 -0.024092 -0.045158) (0.018386 0.072713 0.031665)
Mesh has 3 geometric (non-empty/wedge) directions (1 1 1)
Mesh has 3 solution (non-empty) directions (1 1 1)
Boundary openness (4.26091e-18 2.6356e-17 6.38808e-17) OK.
Max cell openness = 2.66769e-16 OK.
Max aspect ratio = 5.89081 OK.
Minimum face area = 1.03109e-08. Maximum face area = 1.48037e-06. Face area magnitudes OK.
Min volume = 9.38591e-13. Max volume = 4.3971e-10. Total volume = 3.19717e-05. Cell volumes OK.
Mesh non-orthogonality Max: 67.4318 average: 17.5055
Non-orthogonality check OK.
Face pyramids OK.
Max skewness = 0.943828 OK.
Coupled point location match (average 0) OK.

Mesh OK.

End

Would you say this is good enough to run an incompressible pimpleFoam solver in it that has both inflow and outflow (as such I suspect I should be careful about stability)?

[saifee]

Also if you look at http://www.wolfdynamics.com/wiki/tipsandtricks.pdf , on p. 39, it talks about some ranges of non-orthogonality and the number of nNonOrthogonalCorrectors to use as per those ranges.

What do you say what those ranges are actually -- are those angles in degrees or the number of non orthogonal faces in the mesh?

[sri_0548]

Originally Posted by saifee

I tried making some improvements to the mesh and this is the checkMesh results:



Would you say this is good enough to run an incompressible pimpleFoam solver in it that has both inflow and outflow (as such I suspect I should be careful about stability)?

Hi Saifee,


Could you tell me how you got rid of the nonorthogonality?