Contact Interfaces ignored during analyses at random

[DK_9562724]

I am using Creo Simulate in Creo 9.0.3.0

I am having an issue during analysis of structural components using frictionless contact interfaces. The issue is that the defined contact interfaces are occasionally ignored during the analysis process. I have started 2 analyses at the same time with the same mesh and had this issue in only one of them. I have been having this issue in multiple, completely unrelated models.

[rrabe]

Check that the contacts are well defined. Do they have interpenetration? Are the pairs well defined - remember that Creo can be tricky when it comes to selecting surfaces due to the way parts are modelled... Also, a picture or scheme will help the community to give you a better idea...

[DK_9562724]

The contact pairs are well defined surface to surface and they do not have interpenetrations. Here are images the best I can without too much model detail. 

Image with multiple interfaces failing:

Image of the definition of those interfaces:

It is worth reiterating that the issue is random. Sometimes these interfaces work correctly and other ones fail instead, so I don't see how it could be a problem with interface definition.

[ToddKraft · Accepted solution]

DK_9562724,

For issues like this, it is helpful to manually define the output steps.  This gives a higher success rate for the surfaces to connect.  Also putting the surfaces fairly close to each other at the start of the simulation helps.  For troubleshooting, start with Quick Check convergence to ensure the parts come in contact, so you don't have to wait for converged results.  

Regards,
Todd Kraft

Creo Product Manager

[DK_9562724]

Hi Todd, thanks for the idea, I will try some smaller time steps shortly.

Almost every interface in the model is already in perfect contact at the start of the analysis, save for a small clearance around the main pin so I still think it is strange that some of these surfaces would not work. Do you have any ideas what the actual cause may be?

Additionally, in assemblies such as this, I have actually found it to be substantially quicker to iterate by using multiple pass adaptive convergence and checking the result after the first two passes. Interestingly, I have noticed on occasion that interfaces which seem to work on the first pass may fail on the second pass.

[skunks]

please upload step data

[DK_9562724]

Unfortunately I cannot do that.

[DK_9562724]

Hi Todd, I have tried defining steps of a few different sizes now and it has not been successful. The issue persists. Interestingly, now I can see that the interfaces only break after around step 0.2

[ToddKraft]

Hi DK_9562724,

There is a rattlesnake in there someplace.  Without seeing the geometry, we can't help much further to troubleshoot.  Other ideas are to increase the number of elements in the contact area., look for geometry checks in these troubled areas, step geom out and back in (don't like this option, but it sometimes works), degrees of freedom are not making sense to the solver, part accuracies are far different from each other which are in contact, etc... If you can simplify the problem and share, then we can look deeper. 

Regards,
Todd Kraft

Creo Product Manager

[DK_9562724]

Hi Todd, thanks for doing what you could to help with the troubleshooting. I'll see if I can simplify the problem to a point where I can share it, but I think it may be a week or two, unfortunately. 

As for the other suggestions, I have already tried significantly refining the mesh (from around 50000 total elements to over 2 million), I have tried moving some geometry, and the part accuracies are all within less than half of an order of magnitude of each other. As far as I can tell, all degrees of freedom are controlled. The mesh refinement and geometry movement sometimes seemed to lead to different interfaces breaking first, but with no rhyme or reason.