tetrahedral elements exploding in size

[doherts]

Hello,

I've been working on a displacement driven model and am coming across an unexpected behavior where certain tetrahedral elements seem to rapidly expand out of the model. I thought the model would behave well, given the prescribed displacement to constrain the model while a tie is applied. The model is too large to upload to the forum, can I share it over email with someone for tips/advice?

[c.adair.p]

When you say a tie is applied, are you referring to a tied contact interface? If so, I recommend using a large penalty (1000 - 1x10^6). Tied contact requires a much larger penalty to be successful. I also recommend adding the "vector gap" output, which will calculate how much overlap there is between two tied surfaces.

[doherts]

When you say a tie is applied, are you referring to a tied contact interface? If so, I recommend using a large penalty (1000 - 1x10^6). Tied contact requires a much larger penalty to be successful. I also recommend adding the "vector gap" output, which will calculate how much overlap there is between two tied surfaces.

Thanks for the reply. I have tried a tied node on facet surface with auto penalty of 1 and penalty of 1. I have also tried a tied elastic surface, with auto penalty of 1 and penalty of 10000. Both cases have resulted in a situation similar to the below image where certain tetrahedrals explode in size:
image.png

[maas]

Hi doherts,

Can you please share your febio input file? That might allow us to get a better understanding of your setup. You can try to zip the file to make it smaller. If it's still too big, please consider sharing it via a Google drive or other file sharing service.

Thanks,

Steve

[doherts]

Hi doherts,

Can you please share your febio input file? That might allow us to get a better understanding of your setup. You can try to zip the file to make it smaller. If it's still too big, please consider sharing it via a Google drive or other file sharing service.

Thanks,

Steve

Here is a link that should work, I have the boundary conditions and geometry defined in separate files. I have two tetrahedral elements commented out in the geometry file since they seemed to cause convergence issues.

Flat_Tissue.7z

[maas]

Hi doherts,

I got the file. I'll take a look, but can you please clarify something. There is unclosed comment, starting on line 4415 on flat_draft_model_bc.feb.

<!--prescribe bc="y" lc="3">-11.114202</prescribe>
<prescribe bc="z" lc="3">1.17651</prescribe>
<!--<prescribed bc="z" lc="1">1.0</prescribed>-->
<!--<force bc="z" lc="1">10.0</force>-->

So I am not sure if you intended to include the second prescribe boundary or not. Can you please clarify?

Thanks,

Steve

[doherts]

Hi steve,

sorry for confusion, those are outdated boundary conditions. I cleaned up my input files and have uploaded a new copy. As before, I still have tet elements 53927 and 53302 commented out as well:
Flat_Tissue.zip

[maas]

I think I found the problem and the issue appears to be with the mesh. There are a handful elements inside the mesh that are not connected to the surrounding mesh. These elements can undergo unpredictable deformations and be a source of instabilities. The elements I found are: 24698, 53302, 53880, 53927. You can find them as follows:

- open the plot file and click on the element selection button. Make sure to select connected elements.
- click on any element of the skin (or fat) tissue and click 'h' (or menu edit\hide selection). This should select and hide all the elements of the skin (or fat) tissue. Repeat this for the other tissue.
- The unconnected elements are really tiny, so you probably won't see them at this point. To find them, do a region-selection (click+drag with the left mouse button while holding down shift) and cover a wide area on your screen.
- If all goes well, you should see something similar as below, locating the elements thar are not connected to the rest of the mesh.

flat_draft_model_bc.png

You'll have to go back to your original mesh and remove these elements. Of course, this may create tiny holes, so you'll have to check for that. Alternatively, you can go back to your meshing software and see if there is a way to prevent this. Once you removed these elements, please try the simulation again and let us know what you find.

Best,

Steve