a question about refineMesh
 

Hi everybody!
I' like using refineMesh to refime the mesh like below
http://www.cfd-online.com/Forums/att...1&d=1308747850
but the result is like below
http://www.cfd-online.com/Forums/att...1&d=1308747867
what should I do?
Thanks very much!
http://www.cfd-online.com/home/lf/%C...creenshot2.jpg

I have fund the answer, the refineMesh works well.The refined grids just like the first figure, but the paraview's display is not correct， The diagonal lines are of fictitious. I still do not know how to correct it.

Checking the "Use VTkPolyhedron", paraview will show the correct display.

Hi,

Where is located the "Use VTkPolyhedron" in paraFoam ?

Regards,

Stephane.

reply
 

Hi,
If you compile "PV3FoamReader Module" correctly, the paraview's "Object Inspector" looks like figure below ,you can check "use VTKPolyhedron".
http://www.cfd-online.com/Forums/att...1&d=1309311116
but if you failed to compile "PV3FoamReader Module", paraview can still work, the "Object Inspector" looks like figure below ,you can not check "use VTKPolyhedron", the display is not correct, but the grids are ok.
http://www.cfd-online.com/Forums/att...1&d=1309311105

Hi All,

OpenFOAM cannot use meshes with floating nodes, so no the, diagonals are not fictitious. Valid cell shapes recognized by OpenFOAM can be found here. Therefore, to correctly close the mesh shown in the first post, OpenFOAM will create triangles, which is consistent with the diagonals shown in paraview/paraFoam.

Kind regards,

Francois.

how I can refine a specific region?
 

Does anyone know how I can refine a specific region in a cavity for example?
Thanks!

I have the exact OPPOSITE problem ... refineMesh doesn't add the diagonal, so two faces are orthogonal to another single face, which creates problems

Is there a switch somewhere in the refineMeshDict that I haven't seen ???

For whoever stumbles upon this, the above information seems to be incorrect (at least not anymore in OF-2.3.x). I created a small testcase, which consists of two vertically stacked blocks.
The lower block is then "refined" in 2 blocks. I visualised the result in paraview, once using VTKPolyhedra and once without, and further compared the visualisation with a plot in MATLAB, for which I directly exported the data from polyMesh/points and polyMesh/faces.

The result of the Matlab visualisation perfectly agrees with the VTKPolyhedral picture:
Attachment 46569Attachment 46567

The paraFoam version not using VTKPolyhedra:
Attachment 46568

Also regarding the standard shapes linked, those are the ones recognised, but OF also supports arbitrary polyhedra (compare checkMesh output).

I also attach the testcase including my Matlab script for those interested.