Optimal energy-conserving discontinuous Galerkin methods for linear symmetric hyperbolic systems

TL;DR

This paper introduces energy-conserving discontinuous Galerkin methods for symmetric hyperbolic systems, achieving optimal error estimates and demonstrating superior long-time simulation performance on various meshes.

Contribution

The paper develops new energy-conserving DG methods with optimal error estimates for symmetric hyperbolic systems on unstructured meshes, including a high-order energy-conserving time discretization.

Findings

Optimal error estimates of order k+1 in 1D and multi-D on Cartesian meshes.

Method with doubled unknowns is optimally convergent on triangular meshes.

Numerical results show superior long-time performance compared to classical methods.

Abstract

We propose energy-conserving discontinuous Galerkin (DG) methods for symmetric linear hyperbolic systems on general unstructured meshes. Optimal a priori error estimates of order $k+1$ are obtained for the semi-discrete scheme in one dimension, and in multi-dimensions on Cartesian meshes when tensor-product polynomials of degree $k$ are used. A high-order energy-conserving Lax-Wendroff time discretization is also presented. Extensive numerical results in one dimension, and two dimensions on both rectangular and triangular meshes are presented to support the theoretical findings and to assess the new methods. One particular method (with the doubling of unknowns) is found to be optimally convergent on triangular meshes for all the examples considered in this paper. The method is also compared with the classical (dissipative) upwinding DG method and (conservative) DG method with a central flux. It is numerically observed for the new method to have a superior performance for long-time simulations.