CFD Online Discussion Forums

Decomposition methods in parallel

Posted April 23, 2013 at 02:05 by Tushar@cfd (Blogs: OpenFOAM)
Updated March 27, 2015 at 06:51 by Tushar@cfd (Minor corrections)

OpenFOAM has four methods of decomposition:
(1) simple
(2) hierarchical
(3) scotch
(4) manual

"Scotch decomposition requires no geometric input from the user and attempts to minimise the number of processor boundaries."(for more information refer OpenFOAM website)

We tried to compare "simple" and "scotch" methods. These methods were verified for a test case using similar processor inputs:

Both...

Tushar@cfd

Senior Member

Posted in Uncategorized

Views 2372 Comments 1 Tushar@cfd is offline

2D-empty vs pseudo 2D upd. results

Posted April 18, 2013 at 03:53 by skeptik (CFD notes)
Updated April 18, 2013 at 04:15 by skeptik

Pseudo 2D results is here on the graph
Name: kOmega_PisoHexPseudo2D.jpg
Views: 1911
Size: 39.2 KB

Cx is predicted very well, so let's make a comparation with 2D results

Name: table_en.jpg
Views: 1572
Size: 31.4 KB

The difference is really big on supercritical angles, the PISO's results on 2D tetra meshes were also bad and quite similar to the 2D hexa's results.

So, it's better to use PISO with symmPlane or slip BC on domain's sides

#OpenFoam #2Dairfoil

skeptik

Member

Posted in Uncategorized

Views 1336 Comments 0 skeptik is offline

2D-empty vs pseudo 2D - symmPlane

Posted April 15, 2013 at 05:35 by skeptik (CFD notes)

It seems there is big difference in case with 1-cell thick domain where we set empty BC compare to case with 5-cells or 3 - cells thick where we set symmPlane

for SST and NACA23012 we have (17 and 15 grad angle of attack)

Click image for larger version

Name: table.png
Views: 780
Size: 14.5 KB
ID: 137

So when we use PISO, it would be better to use 5 or 3 -cells thick domain.

skeptik

Member

Posted in Uncategorized

Views 1416 Comments 0 skeptik is offline

blockMesh,snappyMesh and experimental results for 2D airfoil

Posted April 1, 2013 at 05:17 by skeptik (CFD notes)

A certain Mahdi wrote me via e-mail, that my tetra-mesh and hybrid mesh was not good for simulations.

So here i show results which were obtained with C-mesh generated by blockMesh utility OF, and snappyHexMesh.

blockMesh domain

Name: hexaMesh1.jpg
Views: 12098
Size: 76.8 KB

Name: hexaMesh1.jpg
Views: 12098
Size: 76.8 KB

mesh near wall

Name: hexaMesh2.jpg
Views: 9190
Size: 88.2 KB

snappyHexMesh domain

Name: snappy1.jpg
Views: 6676
Size: 86.3 KB

boundary layer discretization

Name: snappy3.jpg
Views: 10740
Size: 85.5 KB

...

Attached Images

hexasSimple.jpg (44.1 KB, 4800 views)

skeptik

Member

Posted in Uncategorized

Views 9448 Comments 12 skeptik is offline

about polyDualMesh "Try something smaller than 90"

Posted March 28, 2013 at 02:06 by skeptik (CFD notes)

Quote:

Originally Posted by lillberg

Try something smaller than 90 deg feature angle since it represents the maximum angle between two neighboring surface normals on a surface without corners. That is, the mesh will consider your surface shape a corner if the angel is larger than this value. I typically use something between 45 and 80 depending on the tet mesh resolution on highly curved surfaces. This might be the root of your problem. Looking at my picture the curved surface in the middle would (wrongly) be treated as a corner if

...

skeptik

Member

Posted in Uncategorized

Views 1491 Comments 0 skeptik is offline