Learning Forum

[Mike Youssef]

I have a model of a box with a point mass inside. The box is made up of 6 individual components (sides) and they're all connected using bonded contacts. The geometry of the box is made up of midsurfaces (shell elements). Each side of the box has a chassis slide attached using bonded contacts. The chassis slides are solid elements.

My ultimate goal is to determine the resonant frequency and I've already gotten some results but I wasn't convinced. So I suppressed the boundary conditions and the point mass to check the rigid body modes.

1-3 translation modes are 0 Hz but 4-6 rotation modes are non-zero, in fact they're in the single digit frequency range. So I extracted equivalent total strain and notice non-zero values for strain at each corner of the box. Can anyone shed some light on possible causes? I appreciate any help.

I checked my contacts using the contact tool and don't have any open contacts. There are some that are statused orange.

[peteroznewman]

I assume the point mass is scoped to a Remote Point. Is the configuration of that Remote Point using Deformable or Rigid Behavior?  Try setting it to Deformable.

[Mike Youssef]

Yes the point mass is remote and it is deformable. What appears to have been the culprit is my contacts so I used shared topology instead. Do you have any input on that?

[peteroznewman]

Switching to Shared Topology was going to be my next suggestion.

Another suggestion would be to change the Bonded Contact Formulation to MPC.

[Mike Youssef]

I also tried MPC and didn't get different results. Thanks for your input.

[dlooman]

The most common reason for this is penalty based contact across open gaps or between curved surfaces so it's surprising that switching to MPC didn't fix the problem.  A test is to request stress results and look where stresses are occurring in modes 4-6.  In a rigid body mode there shouldn't be any, so the location of stress could indicate where the artificial model constraint is located.