Preserving Superconvergence of Spectral Elements for Curved Domains via h and p-Geometric Refinement [Slides]
Spectral element methods (SEM) are extensions of finite element methods (FEM) that employ Gauss-Lobatto or similar nodes instead of equidistant nodes for high-order elements. SEM can deliver superior accuracy compared to equidistant FEM due to potential superconvergence. However, significant challenges remain for domains with curved boundaries, which have limited the advantages of SEM for real-world applications. In this work, we propose a novel approach to bolster the overall accuracy and preserve the superconvergence of SEM over curved domains.