Issue when analysing assembly with preload, finite friction contact and large deformation

[rrabe]

Hello all,

I decided to share this after getting, I think, similar issues since Simulate 2.0 and finally finding now a good clue when using Simulate 4.0.

I don´t know if this issue is present in different Creo version and datacodes.

If you use contact with finite friction you must use LDA. If you have the option to not use LDA you may not have this problem.

This simple assembly model (picture) shows a bad combination of features that finishes with:

“*** A fatal error has occurred. *** The nonlinear iteration did not converge for the time value: 0.00000e+00 »

_ What is odd is that this happen only if you are using « single-pass adaptive » and only on the 2nd pass. The 1st pass finished without problem. If you use “quick check” you will not have this problem.

This model has an initial interference to simulate preload and two finite friction contact regions – one between the flanges and one between the head of the screw and the flange.

In short, if “Rigid link” is present – even if it does participate directly in the regions being deformed – the simulation will fail on the 2nd pass (already during preload).

If RL is replaced by weighted-links, the simulation works.

If RL is present and the top contact is suppressed, the simulation works.

If RL is suppressed, the simulation works.

Yes, Creo advices that links and LDA may produce wrong results. However, the above combination looks more like a bug than a wrong result.

So, an old advice in case of odd fails or bad results in LDA – start by suppressing the links when possible.

Hopefully this info can help someone.

Best regards

R. Rabe

[skunks]

"...I decided to share this after getting, I think, similar issues since Simulate 2.0 and finally finding now a good clue when using Simulate 4.0. ..."

-> creo simulate 2 don't have finite friction

[rrabe]

Yes, finite friction was implemented in 3.0 - a welcome implementation!

_ But if you use LDA and contacts you may also run into troubles. This is what I had back in Creo 2.0.

The most common in Creo 2.0 me was to see that sometimes Creo was ignoring some sets of contacts (final contact pressure was being reported as 0). This was not happening when running linear analysis.

Now, from 3.0 on, if you need finite friction you must use LDA.

Would be interesting to verify those troubled Creo 2.0 models again and check if they had RL…

[skunks]

my example for contact+large deformation att. (.mpg)

do you have a example for contact+large deformation?

[rrabe]

Hello

Sorry for the late answer!

A simple model - with two flanges assembled with interference to simulate an initial screw preload (at time 0 - as in the attached picture).

If the contact between the parts is with finite friction you must use LDA - if not, you can choose to use linear or LDA simulation. 

This simple example can reproduce some of the odd results that I wrote about - and even another one that PTC is now aware and hopefully will document (related with time dependence load/unload in LDA with initial preload).

Hopefully this helps!