9. Summary of results

• For a plate with one or more layers, elements DST (QUAD4) give good estimates of displacements, generalized forces, deformations, and stresses (Modeling A and B).

• For a finer mesh, Q4G elements (QUAD4) lead to slightly larger errors, especially for SIXX axial constraints (C and D modeling).

• The two types of elements DST (QUAD4) and Q4G (QUAD4) respectively give the same results with one (Modeling A and C) or two layers (Modeling B and D).

• Elements DST (TRIA3) give much worse results (E and F models). In particular, significant mesh effects are highlighted: the E and F models have the same number of elements but the E mesh is adjusted while the F mesh is free. The error on the arrow is 13% for E and 1% for F. Similar differences in the moments are noted. There is also a strong asymmetry in the moments between the points \(\mathit{A1}\) and \(\mathit{A4}\) for the 2 meshes. This is due to the fact that the meshes are not symmetric with respect to the load. On a symmetric mesh the problem no longer appears. This element therefore has a flexural behavior that is very dependent on the mesh, which undoubtedly merits further study.