After converting an old mesh created using STAR-CD 3.1 to STAR-CD 4.06 by using nastran format, there are holes at particular part of mesh.

Does STAR-CD 4.06 provide any function particularly to repair such meshing problem?

Thank you.

Hi

What do you mean by holes?

Does the mesh has couple? nastran format doesn't support the couple mesh.

Why don't you just read/resume the 3.1 .mdl file directly into 4.06? Nastran does not understand all possible cell shapes that STAR is capable of producing. The holes might be from that.

Hi Allan,

please see the figure http://www.whlam.com/engine/corruptedduct1.gif

There are missing cells at particular region forming some holes.

The reason I choose nastran format is due to the saving problem of old STAR_CD 3.1 .mdl mesh in the STAR-CD 4.06. The resumed STAR-CD 3.1 mesh can not be saved using STAR-CD 4.06. To secure the STAR-CD 3.1 mesh, I therefore exported through nastran format.

It is strange. The problem of holes or more precisely the missing cells still occurs when read the .mdl file directly.

Yes, there are couple settings. Unfortunately I lost the couple settings as well.

Could you share what particular cells nastran format does not support?

The missing cells are trimmed cells which can are not supported within the nastran format.

Should be. The trimmed cell is not widely supported by the standard file format.

In this case, any suggestion to repair it or do you think exporting the geometry and then recreating the entire mesh is a better solution?

Thank you.

NASTRAN does not support trimmed (polyhedral cells). It supports tetrahedron, hexahedron and prism cells.

Here is something you can try. In Star 3.1, use the cdsave command to export the model. In Star 4.06 use the ifile command to read in the model. I'm not sure what happens with the Star 3.1 couples. (Star 4.06 is a face based code which does not use couples.) Star 4 may do the necessary conversion. Sorry but I haven't used Star 4 much so I'm not sure about this matter.

I think exporting the geometry and recreating the mesh is a much better idea. The new trimmed/poly meshers in pro-STAR4 are borrowed from STAR-CCM+ and produce significantly better meshes than older version of STAR3.

Read the mesh into v4 and run the command "cptran". Look if the holes are still there and post the warnings/errors from the output window.

Read the *.mdl file !!!

why not export the *.cel and *.vrt data?

Because even *.cel file format was changed between 3.x and 4.x version.

Many, many thanks to CD-Adapco...

v4 can read the cell files from v3. But with exporting just *.cel and *.vrt you will loose the couples which are there in most proam meshes.

How would you support polyhedral cells without a possibility to write them to the *.cel file? There are also tools available to convert a v4 *.cel file back into v3 format just ignoring the polys.

cptran does seem the right answer to repair the missing cells.

1) I read the v3.1 .mdl file; 2) cptran the mesh (fix the missing cells) ; 3) export as .cel, .vrt and .bnd files (resumed v3.1 .mdl can not saved directly in v4.06); 4) import to v4.06

However, when i check the mesh quality, there are some other error messages.

errornumber-command-description

69-chec- non-right-handed cells found

234-chec- cell has a greater-than-desired aspect ratio

176-chec- cells found with face warpages greater than the prescribed value

80-chec- internal face angle of a cell exceeds prescribed value

70-chec- cells found with negative volume or volume less than specified volume

406-chec- centroidal check failed. cells have centroids which reside outside their volumes.

71-chec- centroidal check failed. cells overlap.

74-chec- there are cracks (unconnected regions) in the model

265-chec- model has groups of disconnected cells.

76-chec- cell faces with very small suface areas found in the model.

Are these fatal errors or ignorable?

What are you doing man? You should simply take the v31 *.mdl file and read it into v406. After that you type the command "cptran all". Then there should be no holes or missing cells.

I tried it myself with an old proam mesh and it worked.

Check if there are any warnings or errors during this process and post them.

Forget the *.cel and *.vrt files!!!

> There are also tools available to convert a v4 *.cel file back into v3 format just ignoring the polys.

BTW, what are these tools? Because sometimes I really miss such an opportunity

I agree to TG. Live the surface from 3.1, then regenerate new trimmed meshes in 4.06. Trimmed cells of v3 and v4 are quite different. Another choice, you may try to resume 3.1 to 3.26, then resume 3.26 to 4.06. But this method may not work...

Ask the support for the file "cell423.f" or search for star-backport.