Self contact in a "plate with a hole" problem.

[Butteran]

Hello!

I am trying compress a typical plate with a hole until it closes to study the effect of the self contact. So far I have been unsuccessful in reaching this state however as the problem is prone to negative system eigenvalues, distorted elements and overall problems with convergence. I am running the problem with a hyperelastic material , for which I have tried models and parameters (mm, MPa,tonne/mm^3), of course with nlgeom on.
To solve it I have tried the following:
- Refining the mesh (both locally and globally) and both triangular and quad elements
- Both explicit and implicit analysis as well as using Riks path following method
- Removing reduced integration and including hybrid elements
- Varying which node i pin to prevent rigid body motion
- Reducing initial/min step size and including line searches
- Both general and self contact interactions

At this point I dont know what to try next as my solutions keep diverging so if anyone has any insight it would be most helpful.
Ps. included an image of how far my simulations get

All the best
Butter

 

[FEA way]

What was the problem with explicit analysis ? Usually it’s used when everything else fails and it rarely crushes since you don’t have to worry about rigid body modes or convergence. If you got severe distortions then maybe it’s necessary to verify the mesh, interactions and loadings (sepcially their magnitudes and rates). Make sure that all values and units are correct and consistent.

Try simplifying the model to see when it fails.

 

[Butteran]

The explicit analysis fails because I am also using periodic boundary conditions (to model a presentative volume element) which causes the analysis to have too many implicit constraints and so It won't run. Basically these bcs are implemented through a set of MPCs and include all faces.

I have verified the grid before running so the elements distort after some point of deformation.
Interaction and load rates are not something I've looked into, might be worth a shot! (forgot to mention that I'm doing displacement controlled loading).

Thanks for all the input!