[qtian]

Dear all,

I did high reynolds number RANS simulation case one month ago and the results seems ok. Now I am trying to do a low reynolds number RANS simualtion with LaunderSharmaKE turbulence model for comparison. I just refined the coarse mesh which I used in the high reynolds number simulation only along the vertical direction in order to resolve yplus=1. For some reason, my simulation for the low reynolds number case has convergency problem and with "unbound k and epsilon" and also has large continuity error. I also tried to turn off the turbulence first and no help also.

I generated mesh in gridgen with fluint format. Here is the results from fluentMeshToFoam and checkMesh. I don't know how to improve these severe non-orthogonality and skewed meshes? Does anyone know how to visualze these severe non-orthogonality and skewed meshes? I don't even know where they are. Thanks for your help.


Default patch type set to empty

Checking mesh...Number of non-orthogonality errors: 0. Number of severely non-orthogonal faces: 149.
--> FOAM Warning :
From function primitiveMesh::checkFaceSkewness(const bool report, labelHashSet* setPtr) const
in file meshes/primitiveMesh/primitiveMeshCheck.C at line 838
Large face skewness detected. Max skewness = 531.581 percent.
This may impair the quality of the result.
105 highly skew faces detected.
Failed 1 mesh geometry checks.
Failed some mesh checks.
done.
Writing mesh... to "constant/polyMesh" done.


checkMesh

Domain bounding box: min = (-0.469 -0.000233522 0) max = (0.78 0.254 0.468) meters.

Checking geometry...
Boundary openness in x-direction = -2.03288e-20
Boundary openness in y-direction = -4.14707e-18
Boundary openness in z-direction = 3.01544e-19
Boundary closed (OK).
Max cell openness = 2.26317e-19 Max aspect ratio = 45.748. All cells OK.

Minumum face area = 4.0582e-08. Maximum face area = 0.00063478. Face area magnitudes OK.

Min volume = 1.66625e-10. Max volume = 2.40859e-06. Total volume = 0.147875. Cell volumes OK.

Severe non-orthogonality for face 1 between cells 0 and 177: Angle = 76.8354 deg.
Severe non-orthogonality for face 526 between cells 175 and 352: Angle = 74.5599 deg.
Severe non-orthogonality for face 3242694 between cells 1086602 and 1086779: Angle = 77.6354 deg.
Severe non-orthogonality for face 3294987 between cells 1104125 and 1104302: Angle = 75.4307 deg.
Severe non-orthogonality for face 3347280 between cells 1121648 and 1121825: Angle = 72.0853 deg.
Number of non-orthogonality errors: 0. Number of severely non-orthogonal faces: 149.
Mesh non-orthogonality Max: 86.6762 average: 8.18947
Non-orthogonality check OK.

Writing 149 non-orthogonal faces to set nonOrthoFaces

Face pyramids OK.

Severe skewness for face 1058 skewness = 5.31581
Severe skewness for face 53351 skewness = 5.30193
Severe skewness for face 105644 skewness = 5.28783
Severe skewness for face 314816 skewness = 5.22916
Severe skewness for boundary face 3550338 skewness = 2.75832
Severe skewness for boundary face 3550339 skewness = 2.56192
Severe skewness for boundary face 3550340 skewness = 2.33748
Severe skewness for boundary face 3550341 skewness = 2.09081
--> FOAM Warning :
From function primitiveMesh::checkFaceSkewness(const bool report, labelHashSet* setPtr) const
in file meshes/primitiveMesh/primitiveMeshCheck.C at line 838
Large face skewness detected. Max skewness = 531.581 percent.
This may impair the quality of the result.
105 highly skew faces detected.
Writing 105 skew faces to set skewFaces

Minumum edge length = 1.54184e-05. Maximum edge length = 0.0312349.

All angles in faces are convex or less than 10 degrees concave.

Face flatness (1 = flat, 0 = butterfly) : average = 0.999998 min = 0.999867
All faces are flat in that the ratio between projected and actual area is > 0.8

Geometry check done.

Number of cells by type:
hexahedra: 1191564
prisms: 0
wedges: 0
pyramids: 0
tet wedges: 0
tetrahedra: 0
polyhedra: 0
Number of regions: 1 (OK).
Failed 1 mesh checks.

[hjasak]

You need to adjust your discretisation on k and epsilon. Do upwind on convection and Gauss linear limited 0.7 on the laplacian. Apart from that you may need to tighten the pressure tolerance.

The above will remove the unboundedness problem and the code should behave OK.

If you want to visualise the bad cells, play around with
subsetMesh, which will allow you to see the bad cells.

Enjoy,

Hrv

[andrewburns]

I almost always have similar problems with k and epsilon, here's what I usually do.

Run a laminar solution for a few hundred iterations, or however long it takes for the residuals to drop to about 1e-3.

Turn turbulence back on, usually it still goes crazy at this point so then...

As Hrvoje Jasak said change the laplacian for k and epsilon to linear limited 0.7 and I also change the relaxation factors on k and e to either 0.5 or 0.3.

This usually works for me.