General Contact (Nonlinear, Hard Contact) in CCX

Sergio

And what about the cpu costs for including more faces in one side of the contact set? I came from old school where you must optimize every element and contact defintion to get the results as soon as possible, that I why I get the procedure of separete the faces.

rwhirley

For tied interfaces the contact search should be done only once at initialization (since the parts do not move relative to each other), and thus should not influence CPU time very much for a nonlinear problem with many iterations or steps. The contact search is typically much more CPU intensive than contact constraint enforcement.

I have not looked at the CCX code to confirm, so be sure this checks with your experience.

Victor

Those are some good points I didn't think about. CCX's node-to-face contact has a SMALL SLIDING option to tell it to search for contact pairs only once per increment. Face-to-face contact (the default, and what Mecway's Contact (CCX) exports as) seems to automatically behave in a "small sliding" way only so it should already not be too slow.

Sergio

Hi, I'm doing some test with hard contact, and I have computed a value for the slope of pressure-overclosure, that make my problem solve. But now I want to start to try with different friction coeficcients, but there is a red warning in the tree telling me that the friction stick slope must be positive for valid friction aplication. Any advice in how to compute this value then?

Regards

Victor

From the CCX manual, it seems to be about an order of magnitude more than the slope of the pressure-overclosure curve. I think I've used the same value for both. That should at least be a useful starting point.