Shell on solid-face test

[lijas]

Test that the dofhandler distributes the dofs correctly when a shell (or beam) element is attached to the face of a solid.

Closes #899

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.86%. Comparing base (821a710) to head (8b5753b).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #926   +/-   ##
=======================================
  Coverage   93.86%   93.86%           
=======================================
  Files          36       36           
  Lines        5313     5313           
=======================================
  Hits         4987     4987           
  Misses        326      326           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[fredrikekre]

Nice that no changes were required haha

[termi-official]

Since I will need this in the not-so-far future, is there a good reference which explains how to couple the elements consistently? And can we add the theory to the docs?

[KnutAM]

Since I will need this in the not-so-far future, is there a good reference which explains how to couple the elements consistently? And can we add the theory to the docs?

I feel that I don't understand the problem/question (i.e. how it could be done inconsistently): if two elements share the same entity (face/edge/vertex), they will be coupled. Of course a write-up of how the dof-distribution works in devdocs/topic guide could be nice!

[termi-official]

E.g. if you have a shell element with rotational and displacment degrees of freedom coupled to a solid element (straight line) for which you just enforce that the displacement is the same, then the shell can freely rotate on the corner of the element, so you also have to put constraints on the rotational dofs.

[KnutAM]

Aha, now I understand the problem, but I don't know the answer.

[koehlerson]

Just don't is the answer 🙃

[KristofferC]

https://youtu.be/Bg2dHz2dxEM?si=H9saoVFW5OJCi5-M might be of some use.