Welcome to the forums. Please post in English or French.

You are not logged in.

#1 2021-06-30 08:57:58

Rethu
Member
Registered: 2021-06-09
Posts: 5

Nonlinear geometry problem with shell and beam elements

Hi!

I was trying to run the enclosed model to try to analyze a large deflection task with nonlinear geometry. My problem is, that no matter how much time substep I define (the enclosed model has 10k substep), I just can't make it converge. Not to mention that the calculation time is way more than expected for a model of this complexity. Could someone please point out, where have I gone wrong? Any help would be appreciated!

So far I tried to increase the number of substeps, but with no result, it only increased computation time, but didn't helped the convergence. The second thing I experimented with is the convergence criteria, but after a limit I didn't want to increase it, I was afraid it would result in an inaccurate result. I also tired to use second order element (TRIA6), but I didn't find any shell element compatible with that (currently I use DKT with TRIA3). COQUE_3D with TIRA7 also didn't give any successful calculation.

Thank you all in advance!

Rethu


Attachments:
plate.zip, Size: 60.37 KiB, Downloads: 17

Offline

#2 2021-07-21 11:59:18

ywan459
Member
Registered: 2020-08-18
Posts: 4

Re: Nonlinear geometry problem with shell and beam elements

Generally speaking the convergence issue should not be a usage problem, but more likely an modelling issue.

There is one thing you could do to improve the rubustness of your solution procedrue. Please try following two statements for the definition of increments in STAT_NON_LINE,

step = DEFI_LIST_REEL(DEBUT=0.0,
                      INTERVALLE=_F(JUSQU_A=1.0,
                                    NOMBRE=200))

timeline = DEFI_LIST_INST(DEFI_LIST=_F(LIST_INST=step),
                          METHODE='AUTO')


It does automatical step cuts. I have tried a trail run. The solver stopped at around 66% due to some internal crash error which has nothing to do with convergence. I am using Windows version, I suppose it could be the root cause. I recommend to try the native Linux binary to see whether works or not. Good luck.

Offline