Showing results for 
Search instead for 
Did you mean: 

How to get convergence with non linear contact


Hello Femap enthousiatics,

I would like to have some best practice advise on the following issue:

I'm trying to simulate the behavior of a ring in a groove that will be pushed out. I used SOL601 (NX Nastran) because of large displacements and loads with different time lines. Contact between the ring and groove is found and I can even let the ring 'grow' by increasing the temperature of the ring. But when I try to push the ring sideways to move it out of the groove, no convergence can be found.

See image: Round steel of 4,5mm diameter is bend to a spring with diameter 100mm. The groove is inside a cilinder, inside diameter 97 mm, groove maximum diameter 100mm. So when the ring needs to be pushed out the groove, is needs to be compressed 100-97 = 3mm in total.

My idea is that the changing diameter of the spring will lead to line of even point to surface contact between ring and groove (Yellow and green in image). 


Can somebody give advice one how to get convergence at the moment the ring starts moving to the right side? Which settings do you prever, and maybe someone can explain why this is so hard to analyse?

Cross section.JPG


Re: How to get convergence with non linear contact


Re: How to get convergence with non linear contact

Solution Partner Creator Solution Partner Creator
Solution Partner Creator

Hello assafwei,


thankyou for your reaction. The first thread is something I already viewed, but I will give changing these parameters an other shot.

The second thread I also viewed and I defined very small time steps (2.000 step for 1 second), this didn't help much. Also used the 'Total Load, Stabilize" option. The results was a converged model, but with only two time steps with results. I need to see the behavior and want to be able to plot the force-way function.

Do you know how to extract information from the run with "Total Load, Stabilize", to use with the more common Auto Increment = 'on'?