Converging forces, but diverging displacement


I'm running a non-linear, static structural model (plasticity, large deflections and frictional contacts), but it won't solve for fine meshes. The model consist of a plate fixed at one end, and bolted to a weight (represented as a solid block with planar face). Forces are applied to the weight. The residual forces are below the criterion, but the displacements are diverging at roughly 70% of max load, and the solution fails due to highly distorted elements. I've verified through hand calculations and a linear analysis that the load is appropriate for the structure. I'm not expecting large plastic deformations at all, just some plastic strain at corners and around bolt holes.

The highly distorted element is close to the bolt hole, and this is where I find the maximum residual force as well. However, there is no contact between this element and the bolt, so I'm surprised to find a large residual here. There is only the frictional contact with the weight, but as seen in the contact status, the status is "near" when the solution fails.

Any thoughts on what could be causing this issue?

Convergence plot (Step 1 is bolt pretension, step 2 and 3 is gradually applying the loading):

Newton-Raphson residual:

Contact status:


  • peteroznewmanpeteroznewman Member
    edited October 2020


    The last image shows the flat face of the part with the hole in it, showing a contact status. Why is the flat face in the contact definition? Try the solution without that face in the contact set.

  • The part with the hole in it (and the rest of the plate) is in contact with the weight (which is hidden in the screenshots). I'm sorry that I can't provide a snapshot of the whole geometry, but I think that would violate company policy. However, I got a simplified version of the structure to converge without problem, so I'll try to build it back up from there!

    I am still curious about what would cause the forces to stay below the criteria while the displacements remain large... In my mind, the displacements are calculated based on the residual forces, so if the residual forces are low so should the displacements be. Do you think it is due to the forces and displacements are averaged over all nodes, or due to the stiffness of some DOFs being very small?

Sign In or Register to comment.