Strange Extrusion (Prism_3D) behaviour. IS IT A BUG?
Strange Extrusion (Prism_3D) behaviour. IS IT A BUG?Posted by Matvej at March 21. 2008Dear developers of SALOME!
Thank You for this GREAT product!
But I have some troubles with SMESH module.
When i'm trying to mesh with extrusion algo coarse mesh (top and bottom surfaces), all is ok, but as mesh density is increased (number of tets in top and bottom surfaces) i recieve OCC_EXCEPTION (SIGSEGV) and volume mesh is not produced (however, surface mesh is normal)
Sorry for bad english
Re: Strange Extrusion (Prism_3D) behaviour. IS IT A BUG?Posted by gunawan at May 20. 2010
Could You help me to mesh my s'tep file, i was imported from solidwork n autocad
my file is here : http://www.4shared.com/file/5NdgLrjv/crane_hook_Gunawan.html
I was try to mesh but I got the problem like the model is not solid.
i hope u can help me to mesh my cranhook in step.
Try mesh algorithms Tet(netgen)/Triangle(mefisto)/Wire discretisation/segments around vertex instead. I've had some different errors though (see my other posts), but those could be related to the base model - as I do have success depending on who creates the base Catia/step model for me. With a part about 1m^2 x 25mm thick I use values for these algorithms of 1000/100/10/(1 or 0.25). The real mesh density, and compute hardware requirements, are most sensitive to the wire discretisation setting, a range I've found from 5 to 20 works with my hardware on this size of part (P4@2.4Ghz, 1GB ram). I seemed to have more trouble with other algorithms. hope that helps. John
If when you try to "show error element" that is in the mesh module and it doesn't show what the problem is (recent issue for me) then I had to find the solution with the following steps...
Try splitting your model into smaller sections. I solved one problem I was having by successive splits of the part and trying to run each side independently (create a big block tool in "geometry" and chop off one side of the part and mesh the other side - you'll find it helpful to keep segmenting these sections and save to a file _before_ meshing, so you can keep using the same cutting tool and not too much is stored in the file).
Test meshing both sides of a split, discarding any Working side (only concerned with Non-working sections). Keep splitting the non-working side until you get a small piece where the problem is and you can focus on that. The small sections allow you to use a refined mesh that you couldn't on the whole model (so see if mesh density is the problem). If you find a mesh size that works (you can turn the problem on/off in the small part) then retest at that working mesh level on the whole part. You may need to change how geometry is constructed or other solutions - but at least this test method will help you isolate where the problem is.
Theory behind this is from Six-Sigma problem solving process called a "dictionary split" - for a parlor game, if you take any dictionary, and have someone pick a secret word, you can find that word in 17 or fewer guesses or "splits" ("is it A-M or N-Z?", then "is it A-G or H-M?", etc, "right or left side of the page", etc).
Activate by Site Administrator on Sep 02, 2009 11:02 AM