×

Notice

The forum is in read only mode.
  1. Forum
  3. Public Forums
  5. Finite Element Modelling - Code_Aster
  7. Convergence problem using contact

Convergence problem using contact

More
11 years 1 month ago #7469 by Joël Cugnoni
Replied by Joël Cugnoni on topic Re: Convergence problem using contact
Don't forget to post the message files.

And please respond by pressing replpy on the previous topic, it will be simpler to follow .
Joël Cugnoni - a.k.a admin
www.caelinux.com
  • Comment
  • Comment's Avatar Topic Author
  • Offline
  • Junior Member
  • Junior Member
More
11 years 1 month ago #7472 by Comment
Replied by Comment on topic Re: Convergence problem using contact
Hello,

So my my calculation runs during approximatively 15 hours this night and stops with the message "manque de temps de CPU" .

15h seems to me already pretty high, should i increase this value in ASTK?

Attached are my files including the .mess,

Thanks


File Attachment:

File Name: Compr_CONTI.zip
File Size:757 KB
Attachments:
More
11 years 1 month ago #7473 by Joël Cugnoni
Replied by Joël Cugnoni on topic Re: Convergence problem using contact
Hello,

it seems that the friction "stick/slip" condition has difficulties to converge... which induces a lot of newton iterations and "point fixe" loops with many time step cut backs. However the system seems well conditionned, so usign newton resolution of contact and friction can speed it up quite a bit. Also, this could be that the "coef_pena_frot" parameter is way too high.

Attached is a modified version, with a lower coef_pena_frot and newton resolution of contact that takes about 5-10 minutes to run.

Then performing a convergence study on coef_pen_frot will tell you if this values is acceptable.

Joel
Joël Cugnoni - a.k.a admin
www.caelinux.com
More
11 years 1 month ago #7474 by Joël Cugnoni
Replied by Joël Cugnoni on topic Re: Convergence problem using contact
Sorry, attachement fail to upload when using Firefox... Chromium does it well however.
Joël Cugnoni - a.k.a admin
www.caelinux.com
Attachments:
  • Comment
  • Comment's Avatar Topic Author
  • Offline
  • Junior Member
  • Junior Member
More
11 years 1 month ago #7475 by Comment
Replied by Comment on topic Re: Convergence problem using contact
Hello Joel,

Cool thanks a lot for your response, now my model runs perfectly :)

But this solution brings me two questions: how do you know if the penalty must be increasing or decreasing? and looking again the doc, i have realized that i didn't excluded the nodes (SANS_GROUP_NO_FR) that are in common with the symmetry and the friction condition but the calculation works without error message, is it normal?

Dany
Moderators: catux
  1. Forum
  3. Public Forums
  5. Finite Element Modelling - Code_Aster
  7. Convergence problem using contact
Time to create page: 0.131 seconds
Powered by Kunena Forum