A pressure-robust embedded discontinuous galerkin method for the stokes problem by reconstruction operators

Philip L. Lederer, Sander Rhebergen

Research output: Contribution to journalArticleAcademicpeer-review

7 Citations (Scopus)

Abstract

The embedded discontinuous Galerkin (EDG) finite element method for the Stokes problem results in a pointwise divergence-free approximate velocity on cells. However, the approximate velocity is not H(div)-conforming, and it can be shown that this is the reason that the EDG method is not pressure-robust, i.e., the error in the velocity depends on the continuous pressure. In this paper we present a local reconstruction operator that maps discretely divergence-free test functions to exactly divergence-free test functions. This local reconstruction operator restores pressure-robustness by only changing the right-hand side of the discretization, similar to the reconstruction operator recently introduced for the Taylor-Hood and mini elements by Lederer et al. [SIAM J. Numer. Anal., 55 (2017), pp. 1291-1314]. We present an a priori error analysis of the discretization showing optimal convergence rates and pressure-robustness of the velocity error. These results are verified by numerical examples. The motivation for this research is that the resulting EDG method combines the versatility of discontinuous Galerkin methods with the computational efficiency of continuous Galerkin methods and accuracy of pressure-robust finite element methods.

Original languageEnglish
Pages (from-to)2915-2933
Number of pages19
JournalSIAM journal on numerical analysis
Volume58
Issue number5
DOIs
Publication statusPublished - 14 Oct 2020
Externally publishedYes

Keywords

  • Discontinuous Galerkin finite element methods
  • Embedded
  • Exact divergence-free velocity reconstruction
  • Pressure-robustness
  • Stokes equations
  • n/a OA procedure

Fingerprint

Dive into the research topics of 'A pressure-robust embedded discontinuous galerkin method for the stokes problem by reconstruction operators'. Together they form a unique fingerprint.

Cite this