Version 1.3.5.1
- 04/12/2022
- by Matej Borovinšek
A new release version of the PrePoMax contains the following FEM features:
- added support for mortar contact
It’s described in CalculiX User’s Guide. This is essentially contact formulation with Lagrange multipliers. According to the documentation, it converges better but needs more time per iteration.
Dear FEAnalyst san,
Thank you for your kindly repry!!
Cannot we set the contact mortar with Prepomax GUI ?
You can use version 1.3.5.1 and find it in the Contact pair options:
Based on some tests I run, it gives way better (smoother) contact stress distributions on the contacting surfaces.
1 Like
Have you noticed any performance/convergence changes like those mentioned in the documentation ? In Abaqus, Lagrange multipliers formulation is known to be more accurate but also worse for convergence and computational cost. I haven’t tested it in CalculiX yet but I’m surprised that the documentation claims that mortar contact converges better. Also, have you noticed any significant differences between each type of mortar contact in your tests ?
I was only testing it in terms of contact pressure distribution. I remember there was a difference in computational time but could not say which was faster.
All right, I’ll do some tests and report here then.
1 Like
The tests confirmed my expectations - Mortar contact has convergence issues in some cases and requires more time (about 1.6-2 times more in analyzed cases) to solve.
Dear FEAnalyst san,Matej san,
Thank you very much for your kindly reply!!
By the way , below mode is better than morter ?
Linear mortar
PG linear mortar
CalculiX User’s Manual advises using Linear mortar and PG linear mortar only if Mortar fails.
FEAnalyst san,
Thank you very much!!
In one of my simulation, I tried to use mortar type of contact. I simply changed surface-to-surface to the mortar. The following errors appear:
Is there some specific needs when using mortar?
It seems that there is a problem with depended nodes in MPC …
I checked CalculiX documentation for awareness (e.g. to be used only with static analysis) but I did not find other specifics to follow.
Thanks in advance for any advice.
Does it work with surface-to-surface contact ? Check where these nodes are located and what constraints/BCs are applied to them.
With surface-to-surface contacts it works, but simulation also fails due to “too many cutbacks”. Actually, it seems too many attempts have been reached.
@FEAnalyst , I will check nodes acc. to your proposal.
I wouldn’t use mortar contact until convergence issues with penalty contact are resolved. As mentioned above, mortar contact can be harder to converge.
1 Like