9. Summary of results

The cases studied are:

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

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

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