A polygonal Reissner-Mindlin plate element based on the scaled boundary finite element method
Anna Hellers, Mathias Reichle, Sven Klinkel
TL;DR
This work develops a polygonal Reissner–Mindlin plate element within a fully discretized scaled boundary finite element method to handle arbitrary polygonal meshes, including non-star-convex shapes. It introduces linear shape functions in both the scaling and radial directions and a dedicated assumed natural strain (ANS) formulation to alleviate transverse shear locking in the thin-plate limit. A two-field variational framework is then used to incorporate three-dimensional material laws while enforcing plane-stress conditions on the weak form, enabling 3D constitutive modeling without Poisson’s thickness locking. Numerical validations across multiple geometries demonstrate robust locking alleviation, mesh-flexibility, and effective integration of 3D material behavior, indicating practical applicability to complex plate analyses.
Abstract
In this work, a polygonal Reissner-Mindlin plate element is presented. The formulation is based on a scaled boundary finite element method, where in contrast to the original semi-analytical approach, linear shape functions are introduced for the parametrization of the scaling and the radial direction. This yields a fully discretized formulation, which enables the use of non-star-convex-polygonal elements with an arbitrary number of edges, simplifying the meshing process. To address the common effect of transverse shear locking for low-order Reissner-Mindlin elements in the thin-plate limit, an assumed natural strain approach for application on the polygonal scaled boundary finite elements is derived. Further, a two-field variational formulation is introduced to incorporate three-dimensional material laws. Here the plane stress assumptions are enforced on the weak formulation, facilitating the use of material models defined in three-dimensional continuum while considering the effect of Poisson's thickness locking. The effectiveness of the proposed formulation is demonstrated in various numerical examples.
