Summary of results
======================

The cases studied are:


.. csv-table::

    "Modeling", "Links"
    "A: 3D, DKT and POU_D_E "," LIAISON_GROUP"
    "B: 3D, DKT, POU_D_E, and DIS_TR "," LIAISON_GROUP; "; LIAISON_ELEM: 3D_ POU, COQ_POU"
    "C: 3D, COQUE_3D, POU_D_E, and DIS_TR "," LIAISON_GROUP; "; LIAISON_ELEM: 3D_ POU, COQ_POU"
    "D: C_ PLAN, 2D_ DIS_TR "," LIAISON_ELEM: 2D_ POU"
    "E: C_ PLAN, 2D_ DIS_TR "," LIAISON_ELEM: 2D_ POU"
    "F: C_ PLAN, POU_D_E "," LIAISON_ELEM: 2D_ POU"



* The mesh is very coarse in elements :math:`\mathrm{3D}` and plate. The test deserves finer modeling, as the results are influenced by how the embedment conditions are described in :math:`O`. Modeling :math:`A` leads to an error of :math:`14\text{\%}` at most,


* however, with good consideration of these conditions, the solution is much better (modeling :math:`B` leads to an error of :math:`3\text{\%}` maximum).


* Comparisons of constraints and efforts give good results (modeling :math:`B`). For shell element :math:`\mathrm{1D}`, the results are very good.