Convergence proplem with friction

STEFANO

Hello community

I'm new to code aster and have previous experience in linear simulations in CATIA V5. In my area, the behavior of plastics is very interesting.

As a substitute for this, I wanted to calculate a non-linear contact clip snapping. Unfortunately, I had to find out that the calculation with friction is not that easy.

Although I have succeeded in calculating this clip connection without friction by using the "DISCRETE" formulation and PENALISATION for ALGO_CONT.

But as soon as friction comes into play I get convergence problems after 5/7 timesteps and increaseing the timesteps make it even worse.

After many try and errors, the meshing of the two contact areas with quadrangle mapping in the size 1mm - 0.5mm has proven for me to be the best.

Is there any way to calculate this simulation completely with friction? Or can maybe the framework conditions be formulated in such a way that the convergence does not break off?

I would be happy if someone can help me.

Best regards

Stefan

Attached file:
CLIP_SNAP.zip, 908523kb

mf

Hello,

this .comm runs, there might be better configurations, though. Convergence in the friction loops is still quite slow,

When bending occurs, you should use 2nd order elements,

Mario.

EDIT: The bad convergence from before occurs when the part snaps back, so another solution for better convergence would be to add small springs to dampen node vibrations/oscillations.

Attached file:
CurrentCase.comm, 3618kb

STEFANO

Hello Mario,

thank you very much for your help.

After so many mistakes, it feels like magic when this simulation runs. :)

I'll take a closer look to the topic REAC_GEOM with the value NB_ITER_GEOM and the other contact parameters you set.

Of course I´ll use a 2nd order mesh for the right simulation, the 1st order mesh is only to find out how

this contact simulation works and to save computing time.

I understand what you want to tell me about the vibrations/oscillations of the parts.

It needs the springs to compensate this behavior like a damper. I´ll have a look for an informative test case for this topic.

I also tried out of interest to perform the simulation with several time steps (70 timesteps), the simulation was also well

done but ended unfortunately with "EXECUTION_CODE_ASTER_EXIT_2685=0" as you can see in the message file.

I have an error in NOM_CHAM SIEF_NOEU instead of SIEQ_NOEU but with 7 time steps this simulation ended successfully and that cannot be the reason.

It may have something to do with this error message:

Criterion of destruction of the file (1.00%) associated with the base VOLATILE exceeded 10.00%

Nombre of records used: 6289

But I can't figure it out.

Best regards

Stefan

Attached file:
70_timesteps_code_2685=0.txt, 1213040kb