Model stability

[john_erichsen]

I have a steel model that has two clamp assemblies that in contact with a pipe.   The clamp assemblies are held to the pipe by bolt pretension.   I am running static stress with linear material models.   The contact surfaces are set as surface with friction.   I have added 3d springs to the contact surfaces to stabilize the model.   However, I continue to get a "warning: quasi-periodic solution achieved" which I have interpreted to represent a model result that is incomplete.   If I continue to raise the stiffness of the 3-D springs, at what point do the springs begin to influence the results?   How do I review the model to understand the impact of the higher spring constant?  I have reviewed the deflected shape to confirm the results make sense, however, this seems to be less than an exact approach.

Thank you

[AstroJohnPE]

Hi John,

You may want to try a different "Setup > Model Setup > Parameters > Contact > Iteration Method" (or something like that) to see if that avoids the periodic solution.

My approach is to check the force in the springs ("Results Contours > Other Results > Element Forces > Axial Force"). If the axial force is too high, you need to reduce the stiffness.

[john_erichsen]

I have been using sparse. Is iterative better?

The same warning occurs with a very large stiffness.

So it is the solver or a modeling issue I have not discovered. Thanks

Sent from my iPhone

Thank you
John R Erichsen
10321 Adams Rd
Granger, IN 46530

574-277-0670 x701

[AstroJohnPE]

Hi John,

I think the choice between the sparse and iterative solver does not have an effect on the quasi-periodic result that you are getting with the gap/contact elements. The sparse versus iterative solver relates to how the system of equations is solved. The contact iterative method that I mentioned before is related to how the algorithm adjusts which nodes are in contact and which are not in contact depending on the previous solution. The quasi-periodic result indicates that the same set of nodes were found to be in contact during two different contact iterations, or in other words, the algorithm is in a repeating loop.

But in general, I let the software choose which solver -- sparse or iterative -- to use.