STAR-CCM+: Mixed emotions
 

Hello group,

I am a experienced user of STAR-CD and I started to consider STAR-CCM for some applictions a few weeks ago. I want to express my impressions and I am happy about some feedback from you in comparison of STAR-CD (v3.2x/v4) and ccm+.

In my eyes surface tools are best of these from cd-adapco. Nevertheless, automatic volume meshing is horrifying: I get lots of bad cells with both Poly and Trimmer in 2.08.004 even on perfect surfaces. In my eyes this is not yet stable enough. Maybe I should mention I have very complex geomtries.... The ccm+ solver is good. I think for most cases it is already best from cd-adapco. It is very robust, but I have not yet validated the results. It sometimes (but not always) runs on these very dirty meshes which you normally can not calculate on... I do not know how they do this.

Overall: For me it is still the way to create regulare meshes (not more than 6 faces) and maybe use ccm+ to solve them instead of one of he star-solvers. What are your experiences?

BastiL

Re: STAR-CCM+: Mixed emotions
 

Concerning meshing complex geometries - as far as I've noticed, usually bad cells come from bad subsurface, so this problem dates back to ProAM actually.

Solver is "too robust" sometimes, i.e. solution doesn't diverge but stay at very high level of residuals.

Re: STAR-CCM+: Mixed emotions
 

I also get bad cells without BLs (so no subsurface is needed). You see that polyhedral definition fails. The poly mesher has areas where chnge from Tetra to Poly fails (sharp angled tets looking out of a poly e.g.) and the Trimmer makes large and ondefined self intersecting or concave cells,... I did not have this with the old trimmer of proam. So polyhedral definition makes life more difficult in tis case.

Mayb solver is "too robust" - but in my eyes it is ahead of both both STAR-Solvers concerning robustness and most of the models.

Re: STAR-CCM+: Mixed emotions
 

I am also an experienced STAR user from WAY back â€" trying to become an experienced CCM+ user. I've also noticed that CCM+ will make some bad cells, both trim and poly. However, after I 'Remove Invalid Cells' my models usually run â€" an extension of the concept from ESUhood and ESAero. The exception was where I had some poly cells with too much 'cell skewness'. This came from bad sub-surface surface mesh. Sadly I could view the cells but I couldn't find a way to remove them. If I would have resurfaced the imported surface mesh, I probably wouldn't have had this problem. Here, I'm in agreement with Anton â€" bad volume mesh comes from either bad surface mesh or inappropriately sized volume grid for the geometric detail you wish to capture. In any case, CCM+'s trimmer is way faster than STAR v3.26's ammbatch and I believe that it makes much higher quality cells/fewer bad cells.

Re: STAR-CCM+: Mixed emotions
 

I agree that trimmer is faster. However, in my eyes mesh quality is worst, except from the fact that there are no more unresolved cells. Simply export mesh as *.ccm to STAR v4 and run concave or centroid check....

Re: STAR-CCM+: Mixed emotions
 

Maybe I should add that my geometries are very special: Lots of thin-walled guiding elements .....

Re: STAR-CCM+: Mixed emotions
 

Hi all,

after running ferst larger models I have to add that solver is not only robust but also highly efficient.

Re: STAR-CCM+: Mixed emotions
 

Along with many other enhancements, subsurface generation around sharp corner is significantly improved in STAR-CCM+ V2.10.

http://www.cd-adapco.com/press_room/...tarccm210.html

Re: STAR-CCM+: Mixed emotions
 

Dear Stephen,

thanks for this hint, I will look at it. However I checked my meshes without boundary layers and I also found lots of bad cells in there for both trimming and poly. So this is not only a subsurface topic.

BastiL

Re: STAR-CCM+: Mixed emotions
 

Ok, but what do the quality checks in CCM+ say about your mesh; face validity, cell quality, and volume change? And, does it run?

Re: STAR-CCM+: Mixed emotions
 

Quality checks in ccm+ are always "ok". However. removing bad cells always makes sense. Afterwards the mesh sometimes runs and sometimes not. There is no hard criterion. In general creating finer meshes is helpful. What I wanted to point out with the quality checks is that meshes are definitely not clean and if you can run on these this must have to do something with solver "tricks".