Abstract
We investigate the application of the discontinuous Petrov-Galerkin (DPG) finite element framework to stationary convection-diffusion problems. In particular, we demonstrate how the quasi-optimal test space norm improves the robustness of the DPG method with respect to vanishing diffusion. We numerically compare coarse-mesh accuracy of the approximation when using the quasi-optimal norm, the standard norm, and the weighted norm. Our results show that the quasi-optimal norm leads to more accurate results on three benchmark problems in two spatial dimensions. We address the problems associated to the resolution of the optimal test functions with respect to the quasi-optimal norm by studying their convergence numerically. In order to facilitate understanding of the method, we also include a detailed explanation of the methodology from the algorithmic point of view.
Original language | English |
---|---|
Pages (from-to) | 2096-2113 |
Number of pages | 18 |
Journal | Computers and Mathematics with Applications |
Volume | 66 |
Issue number | 10 |
DOIs | |
State | Published - Dec 2013 |
Externally published | Yes |
Keywords
- Automatic stabilization technique
- Convection-dominated diffusion
- Discontinuous Petrov-Galerkin
- Finite element method
- Numerical stability
- Robustness