Dear all,

I'm trying to simulate the crumpling of a hollow cylindrical crash box by a dynamic explicit study using abaqus cae. The material of the box is defined as mild steel and a front plate of 250kg was made to get in contact with the crash box at 11ms-1. The stable time increment was 2.66e-7 at the beginning. The study failed due to excessive rotation of 3 nodes after 133914 increments. However in results some elements were seen to pierce through the front and back plate which were defined as discrete rigid. Could someone help me figure out the reasons for this to happen?

At interactions self contact was created only to the crash box using penalty contact method and defining friction as the interaction property.

Tie constraints were used on both the top and bottom plates to fix them with the crash box.

Boundary conditions were created to fix the bottom plate in space and to let the top plate move only along the axis of the crash box.

The crash box was meshed using free quads with an element size around 3mm.