Collet Design - Surface Contact Problem

[tfjield]

I'm evaluating the forces and stress in a collet-type design, and I'm having problems with the surface contact.

In the image below, the purple tube on the right of the image moves to the left with a specified amount of force, causing the collet fingers (green) to close on the blue rod.  I've specified a sliding surface contact between the purple tube on the right and the green collet fingers, and that appears to be working properly.

Unfortunately, the fingers move through the blue rod as if it's not even there.  I've tried specifying a variety of surface contacts, but none of them prevent the fingers from moving through the rod.  Perhaps it's because of the angles of the surface?  Any ideas?  Thanks in advance.

[AstroJohnPE]

Hi,

The first step is to look at the log file and confirm that each surface contact pair is listed. Just before it starts the iterations, it lists all of the parts, then each of the contact pairs. There should be one entry for each contact pair that you defined. (The initial number of contact elements may be listed as zero, but I think that in itself is okay.)

Next step is to tweak the contact settings. I think that MES limits the distance at time zero in which it will consider contact elements based on the mesh size. So if the mesh is small compared to the gap between the parts, this could have an effect. (Shoubing will need to add some details here, especially if I am wrong!) Right click on a contact pair in the browser and choose "Settings". When the dialog appears, click the "Advanced" button. When the new dialog appears, click the second tab ("Geometry" if my memory is correct) and change the first input (something "contact interaction distance") to a value that is larger than the initial gap.

If the fingers are moving so fast that they go from not being in contact on one time step and then completely pass through on the next time step, the previous setting can help with that, too. Alternatively, use a smaller time step or more gentle force to bring the fingers in contact with the bar, and then ramp the force up to full load.

[tfjield]

Thanks, John.  Am I correct in assuming that I need to run this as a Static Stress with Nonlinear Material Properties?  I do get the movement if I do a linear model, just not the appropriate surface interaction with those two parts.

[AstroJohnPE]

Yes, I think you need to use the "stress with nonlinear materials" because it looks like the parts are not in contact at time 0, and linear stress only creates contact elements where the parts are in contact at time 0.

If you use a prescribe displacement to move the purple part, then Static Stress with Nonlinear Materials may work. If you apply a force to a part that is "free" to move in a static analysis, the solution is that the part accelerates to infinity over an infinite time (infinite time is the assumption in static stress). But if you use MES with Nonlinear Materials and apply a force, the inertia of the body prevents it from accelerating to infinity, so it should work.