*WARNING in gentiedmpc: no tied MP

Hello,
I am working on a project and I constantly receive warnings like this while doing analysis. Stress is also very high. There’s definitely a mistake somewhere.
image

It means that some nodes couldn’t be connected with tie constraint, likely due to a significant mismatch in meshes on both sides. Try increasing the tolerance and make sure that adjustment is enabled.

You could also try creating a compound part instead of connecting the regions with tie constraints if that’s ok for your workflow.

Unfortunately, the composite piece is not okay for my project. How much should I increase the tolerance?

Or what else can I do?
Thanks a lot!

The tolerance setting should correspond to the largest orthogonal distance between the slave node and opposing master face that should still be connected plus some small distance as a reserve. Don’t forget about the adjustment.

Unfortunately I still get the same warning. What could be another solution?

Make sure that master surface is the one belonging to the part with coarser mesh. You should check where tie fails to connect the surfaces and possibly reduce the mismatch between the meshes.

You could also replace the tie constraint with tied contact.

Could it be because of this warning that it has high stress values?

Tie constraints in general can give spurious stress concentrations in some cases. They should be avoided near the regions of interest (same with most types of constraints).

Did you try with tied contact ? It may give better results in this case.

And it may also matter if it’s 2D or 3D in CalculiX when it comes to tie constraint.

I tried it with tied contact and the result is the same. I tried everything you said and I still can’t figure out what the problem is.

With proper master/slave assignment ? What about mesh adjustment ?

Can you share the .pmx file ?

Is it possible that two close nodes are merged?

You can upload it to some hosting website like WeTransfer, Google Drive or Dropbox and paste the link here.

You have to set access to “Anyone with the link” before copying the link and pasting it here.

Try a frequency analisys. You will find the problem instantly.

Your geometry is much more complex than I expected based on the screenshot. It looked like 2D.

Anyway, the first tip is that it would be better to use hex or hex-dominated meshes. Most parts are simple enough for that and they are very thin so even shells (with some additional solids if needed) could be the way to go.

When it comes to connections, it seems that there are only tie constraints and contact pairs applied to the same surfaces. That’s how you tried to model tied contact ? It should be done by suppressing tie constraints and using only contact pairs with proper interaction (surface behavior type set to Tied). Also, there are no connections between the long sides of the 3 parts. Should they be connected only by the 4 C beams underneath ?

Yes, there is no connection between the long edges of the 3 pieces. It should be like this.

How can I use hex or hex-dominated meshes?

There’s a series of YouTube tutorials about that. This one is about extruded meshes that could be used here if the parts were divided into more segments: https://www.youtube.com/watch?v=TX-YtD7VOmc

But this would make it necessary to add even more connections so maybe let’s just focus on fixing the issues with existing ones first. Or try using transfinite mesh for properly partitioned (divided into basic volumes) parts: https://www.youtube.com/watch?v=veFObRL8cQY

1 Like