A priori analysis of global and local output error estimates for CG, DG and HDG finite element discretizations

Abstract

In this thesis, a priori convergence estimates are developed for outputs, output error estimates, and localizations of output error estimates for Galerkin finite element methods. Specifically, Continuous Galerkin (CG), Discontinuous Galerkin (DG), and Hybridized DG (HDG) methods are analyzed for the Poisson problem. A mixed formulation for DG output error estimation is proposed with improved convergence rates relative to the common approach utilizing statically condensed, p-dependent lifting operators. The HDG output error estimates are new and include the impact of stabilization. Comparisons to numerical results demonstrate (1) the sharpness of the estimates and (2) that the HDG estimates are approximately an order of magnitude more accurate than CG and DG.