[cyln]

Hello,

I am running a URANS simulation using pimpleFoam solver along with kOmegaSST turbulence model. My simulation has been running without any issues, however, I wanted to update my mesh. I kept all BCs and discretization schemes the same, but now I am having a divergnce issue with the new mesh. The properties are quite close. Below are the DETAILED mesh qualities (checkMesh -allGeometry). Can you notice any problematic issue between two meshes?

This mesh is converging

Mesh stats
    points:           10818179
    faces:            53648093
    internal faces:   53421359
    cells:            22300424
    faces per cell:   4.80123
    boundary patches: 3
    point zones:      0
    face zones:       3
    cell zones:       2

Overall number of cells of each type:
    hexahedra:     6785013
    prisms:        4063020
    wedges:        0
    pyramids:      234710 
    tet wedges:    0 
    tetrahedra:    11217681
    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                   Bounding box
    FARFIELD            64292    32260    ok (non-closed singly connected)   (-3.6 -1.89 -1.638) (5.4594 1.89 1.638)
    OUTLET              6172     3200     ok (non-closed singly connected)   (-3.6 -1.89 -1.638) (-3.6 1.89 1.638)
    BODY                156270   78137    ok (closed singly connected)       (0 -0.105 -0.091) (0.3033 0.105 0.091)

Checking geometry...
    Overall domain bounding box (-3.6 -1.89 -1.638) (5.4594 1.89 1.638)
    Mesh has 3 geometric (non-empty/wedge) directions (1 1 1)
    Mesh has 3 solution (non-empty) directions (1 1 1)
    Boundary openness (-4.75074e-16 5.6058e-16 -5.16538e-16) OK.
    Max cell openness = 6.56332e-15 OK.
    Max aspect ratio = 234.03 OK.
    Minimum face area = 2.54377e-09. Maximum face area = 0.0100949.  Face area magnitudes OK.
    Min volume = 5.77004e-13. Max volume = 0.000334306.  Total volume = 86.5231.  Cell volumes OK.
    Mesh non-orthogonality Max: 78.2651 average: 13.3015
   *Number of severely non-orthogonal (> 70 degrees) faces: 12.
    Non-orthogonality check OK.
  <<Writing 12 non-orthogonal faces to set nonOrthoFaces
    Face pyramids OK.
    Max skewness = 2.65922 OK.
    Coupled point location match (average 0) OK.
    Face tets OK.
   *Edges too small, min/max edge length = 4.95773e-06 0.183281, number too small: 304584
  <<Writing 304584 points on short edges to set shortEdges
    All angles in faces OK.
    Face flatness (1 = flat, 0 = butterfly) : min = 0.924377  average = 0.99978
    All face flatness OK.
    Cell determinant (wellposedness) : minimum: 6.73684e-20 average: 2.86146
 ***Cells with small determinant (< 0.001) found, number of cells: 1151198
  <<Writing 1151198 under-determined cells to set underdeterminedCells
    Concave cell check OK.
    Face interpolation weight : minimum: 0.0767751 average: 0.464583
    Face interpolation weight check OK.
    Face volume ratio : minimum: 0.047799 average: 0.875447
    Face volume ratio check OK.

Failed 1 mesh checks.

End

Below is the mesh diverging

Mesh stats
    points:           12257873
    faces:            62999283
    internal faces:   62781421
    cells:            26552434
    faces per cell:   4.73707
    boundary patches: 3
    point zones:      0
    face zones:       3
    cell zones:       2

Overall number of cells of each type:
    hexahedra:     7746063
    prisms:        3816332
    wedges:        0
    pyramids:      262510 
    tet wedges:    0 
    tetrahedra:    14727529
    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                   Bounding box
    FARFIELD            64841    32535    ok (non-closed singly connected)   (-3.6 -1.89 -1.638) (5.4594 1.89 1.638)
    OUTLET              6239     3234     ok (non-closed singly connected)   (-3.6 -1.89 -1.638) (-3.6 1.89 1.638)
    BODY                146782   73393    ok (closed singly connected)       (0 -0.105 -0.091) (0.3033 0.105 0.091)

Checking geometry...
    Overall domain bounding box (-3.6 -1.89 -1.638) (5.4594 1.89 1.638)
    Mesh has 3 geometric (non-empty/wedge) directions (1 1 1)
    Mesh has 3 solution (non-empty) directions (1 1 1)
    Boundary openness (-3.45339e-17 -2.10547e-16 -6.58183e-17) OK.
    Max cell openness = 7.26584e-15 OK.
    Max aspect ratio = 232.805 OK.
    Minimum face area = 2.35453e-09. Maximum face area = 0.0104033.  Face area magnitudes OK.
    Min volume = 5.0309e-13. Max volume = 0.000305801.  Total volume = 86.5231.  Cell volumes OK.
    Mesh non-orthogonality Max: 76.6128 average: 12.7739
   *Number of severely non-orthogonal (> 70 degrees) faces: 41.
    Non-orthogonality check OK.
  <<Writing 41 non-orthogonal faces to set nonOrthoFaces
    Face pyramids OK.
    Max skewness = 2.64733 OK.
    Coupled point location match (average 0) OK.
    Face tets OK.
   *Edges too small, min/max edge length = 4.73886e-06 0.172407, number too small: 281115
  <<Writing 281115 points on short edges to set shortEdges
    All angles in faces OK.
    Face flatness (1 = flat, 0 = butterfly) : min = 0.939085  average = 0.999831
    All face flatness OK.
    Cell determinant (wellposedness) : minimum: 6.59292e-20 average: 2.78323
 ***Cells with small determinant (< 0.001) found, number of cells: 1130274
  <<Writing 1130274 under-determined cells to set underdeterminedCells
    Concave cell check OK.
    Face interpolation weight : minimum: 0.0722425 average: 0.465362
    Face interpolation weight check OK.
    Face volume ratio : minimum: 0.0518655 average: 0.877726
    Face volume ratio check OK.

Failed 1 mesh checks.

End

[pete20r2]

Well you are failing the checkMesh in both cases for a start (did you read the logs??). Both give short edge errors and the second gives high non-orthogonality errors and small determinate errors.
You might also want to try the (fairly new) residuals objectFunction
https://www.openfoam.com/releases/op...esidual-fields

With this you can write out a few timesteps and see where the divergence is occuring. You can also visualise the bad sets from checkMesh to find where they are (my money is on them being in the same place).

[cyln]

Hello Peter,

Thanks for your reply. When I use checkMesh only, the mesh does not fail and does not give any fail output. I posted "checkMesh -allGeometry" on purpose to provide more detailed information about the mesh.

For the diverging case, the case diverges at the first time step. While both meshes have similar qualities, I wonder why this divergence occurs for the second mesh. What is so different? I am using OF4.1 so not sure if I can use this utility.

[pete20r2]

Are you initialising the fields ie with potentialFoam? Are you under-relaxing the first few steps for stability?

[cyln]

The flow is initialized with the mappped results from the simulation using the first mesh. This simulation has been running for about 2000 time steps without any issues.

pimpleFoam solver already uses under-relaxation in its algorithm, so do I.

[pete20r2]

Can you post any more info, case folder, solver logs etc?
There's not much to tell from just the checkMesh