The
paper provides an
evaluation of accuracy of the DST-BL triangular 9 degrees of
freedom element of Batoz and
Lardeur. The local assessment of
convergence is
performed with using J. Rakowski’s semi-analytical method. The results are in
general negative: the element turns out to produce non-convergent sequences of
approximate solutions, which is especially drastically seen for orthotropic
plates. The best results should be expected for clamped isotropic plates
discretized with a regular hexagonal mesh. Moreover, a question of correctness
of DKT approximation scheme is revisited (the convergence proof by F.Kikuchi
being still valid) thus confirming the convergence for thin clamped isotropic
plates.