General CAELinux/Salome Questions
- Joël Cugnoni
-
- Offline
- Moderator
-
Concerning convergence there are 3 different kind :
1. Convergence of the residual of an iterative linear solver (for example CG)
This can be seen in aster output with info=2. Direct solvers give you directly residuals <1e-8.
2. Convergence of the nonlinear solution or segregated multiphysic coupling(comsol, Elmer)
it can be seen in non linear solutions only , in aster it will be printed in message file
3 mesh convergence: some software can refine the mesh automatically and plot a mesh convergence graph
Although you can do adaptive meshing with Homard, you have to generate such graph manually for a. quantity if interest (max stress?)
Cheers
Joel
Joël Cugnoni - a.k.a admin
www.caelinux.com
- Brian Hale
- Topic Author
- Offline
- New Member
-
- Posts: 7
- Thank you received: 0
I think i have a good 50/50 split of bulbs on to bulbs off at this point

... I get a warning stating "some nodes are not placed on geometry...." when I check "medium nodes on geometry"? .
This is reported by Salome?
Yes, this is reported by Salome.
the <CALCULEL_7> error I kept getting....
This is reported by Code Aster? So why should these errors/remarks be related?
Yes, this is reported in the Code Aster files that are created after I hit run. I am assuming they are related because when I put the entire error into google translate (sorry dont know french) it says something about the mesh being distorted a gauss point and that is messing up a Lagrangian (that part is way over my head).
The bigger reason I felt they were related was if I didn't convert the mesh to quadratic, then I don't get that error.
The short answer is: yes.
I too like short answers but I appreciate you taking the time to give me the long answer!

And I believe you are correct, you only get the plot when doing a non linear in Ansys as well. Its been a while since I used Ansys and I had forgotten that part.
- Brian Hale
- Topic Author
- Offline
- New Member
-
- Posts: 7
- Thank you received: 0
First the sappy stuff....
I am learning all this great stuff on your CAELinux 2013 version you just released, and am I really enjoying it so far. Thank you for your hard work on this project. You've done a great job.
Now the fun part...
1. Convergence of the residual of an iterative linear solver (for example CG)
This can be seen in aster output with info=2. Direct solvers give you directly residuals <1e-8.

Sorry I think i missed this chapter in the tutorials? What do you mean when you say "output with info=2" and "Direct solvers give you directly residuals <1e-8"?
Shiny2. Convergence of the nonlinear solution or segregated multiphysic coupling(comsol, Elmer)
it can be seen in non linear solutions only , in aster it will be printed in message file
3 mesh convergence: some software can refine the mesh automatically and plot a mesh convergence graph
Although you can do adaptive meshing with Homard, you have to generate such graph manually for a. quantity if interest (max stress?)
When you say "some software" do you mean commercial or as part of CAELinux?