Interference-fit analysis

theCapitain wrote:

I'm just doing tests on interference fitted assembly and I started this tread after obtaining curious results using quadratic elements (changing the mesh in Aster). Actually I haven't understood what's the matter.

Hello,

If you attach your latest files *.mesh, *.comm, etc. that would help.

Regards,
JMB<br /><br />Post edited by: JMB, at: 2010/08/29 00:37

Code Aster does not automatically rehash the time step in the case of non-convergence. You can enable it. Check out this thread

www.code-aster.org/forum2/viewtopic.php?id=13528

JMB wrote:

If you attach your latest files *.mesh, *.comm, etc. that would help.

You’ll find the .med and command files on the attachment. Be aware that because of my laziness the Mesh is still referred as Mesh_lin on the command file, although already converted into quadratic elements.
Regards.

theCapitain wrote:

You’ll find the .med and command files on the attachment.

Thanks. I tried to run it with some minor changes to comply with Ver10.2+ but it does not converge! I have attached the mess file here and hope somebody can take a look at it and tell me what needs to be fixed.

Regards,
JMB

One of the problems that this has is that it does not rehash the time step if it doesn't converge. Occasionally in non-linear analysis you get to a point where it doesn't converge because the node penetrates too much or the displacement is too great. When this happens you need to give it a smaller timestep so that the penetration or displacement is less. Code-Aster can do this automatically. See the link in my previous posst for ab explanation in how to do it.

Peter

Peter Halverson wrote:

When this happens you need to give it a smaller timestep so that the penetration or displacement is less. Code-Aster can do this automatically. See the link in my previous post for ab explanation in how to do it.Peter

Hello Peter Halverson,

Thank you for your reply and the explanation. I am sorry I was not very explicit in explaining the situation. I am aware of the cutting back of time (load) steps and have used automatic time (load) stepping often in contact and other nonlinear problems since it has become available recently.

However what puzzles me is a problem that solves in an earlier version with manual load steps fails in the newer version. Furthermore even when I reduced the load steps 5 fold it did not converge. My question is what has regressed? Or am I making some other mistake? I will try the auto load stepping shortly. Thank you for your help and I will keep you and the forum informed.

Regards,
JMB