Thanks for your reply!
I have activated internal pipe contact, but that was not the issue. I found out just now that the 'Time Step Controls' helps. It shouldn't be 'None'. I think this helps to increase the number of substeps and prevents the model from missing the contact when it deforms.
This feels similar to what we do in APDL, we can increase the number of substeps to help identify the contact and ease the convergence. Therefore, I have a further question regarding this. How can I adjust the number of substeps manually in Mechanical? Is there something similar to the 'nsubst' in APDL?
Besides, I found the Time Step Controls option doesn't always help. There are two situations:
- If I don't activate Fz, but only Fy, the ring will still fall off.
- If I turn off Small Sliding, the ring will also fall off when Fy applied. However, I do need large sliding in X-direction. If I don't turn off Small Sliding, the ring will have unrealistic deformation when Fx or Ux is applied at the end of the pipe.