Rigid Body Contact Issues

[aaron_s_fox]

Hi FEBio Team and Users,

I am a relatively new user and have run into some trouble in appropriately modelling contact between two rigid bodies. I have two rigid bodies - a sphere and a concave surface - and am prescribing rigid displacements and rigid forces on the sphere while the concave surface remains fixed. Effectively the sphere will be displaced to just make contact with the surface, and be compressed against it while it is translated perpendicular to the surface by a force. I think I have a good enough understanding of how to apply the rigid constraints across steps to produce this behaviour - however my issue is with the contact. I have tried a number of contact models (rigid, sliding etc.) and none seem to stop the sphere from simply penetrating through the convex surface (i.e. if I displace it towards the surface the it just passes straight through). Is there any specific advice for this problem?

I have also encountered the same issue that presented in the recent ESB webinar, whereby the optimiser just prints out all zeros instead of actual values. In the webinar this occurred in the demo and then Steve opened up a new project - and I don't recall seeing what the cause of this problem was?

Thanks,

Aaron

[ateshian]

Hi Aaron,

When using contact between rigid bodies, you should turn off auto-penalty and Lagrangian augmentation. The penalty now represents the stiffness of the rigid body at the contact interface. This should work with (at least) the sliding-elastic contact interface.

Regarding the webinar issue, Steve will need to respond to that.

Best,

Gerard

[aaron_s_fox]

Thanks Gerard - that works perfectly

For a basic sphere-plane interaction I had to set the penalty quite high to stop any penetration. Presumably this is just requires a bit of experimentation depending on the size and how fine the meshes used are - I doubt there is a hard rule for what the penalty needs to be set to to avoid this?

Aaron

[ateshian]

Hi Aaron,

I am glad that worked. The value of the penalty can be estimated from the Young's modulus E of the materials you are idealizing as rigid bodies, along with the maximum penetration g you would be willing to tolerate between your contacting surfaces under the loads applied in your model. For example, if E = 200 GPa and g = 0.2 mm, you should set the penalty to 1000 GPa/mm.

Best,

Gerard

[maas]

Hi Aaron,

Regarding the zero values during the webinar, unfortunately, I didn't save the project file that I was working on, but from looking at the feb file that was written, it looks like I did not specify material parameters for the neo-Hookean material: both E and v are zeroes. I meant to set these to E =1 and v = 0.3. After loading the feb file back into FEBioStudio and making this change, I was able to run the model successfully.

Cheers,

Steve