David A. Kopriva,Gregor J. Gassner,Jan Nordström
Abstract:We use the behavior of the $L_{2}$ norm of the solutions of linear hyperbolic equations with discontinuous coefficient matrices as a surrogate to infer stability of discontinuous Galerkin spectral element methods (DGSEM). Although the $L_{2}$ norm is not bounded by the initial data for homogeneous and dissipative boundary conditions for such systems, the $L_{2}$ norm is easier to work with than a norm that discounts growth due to the discontinuities. We show that the DGSEM with an upwind numerical flux that satisfies the Rankine-Hugoniot (or conservation) condition has the same energy bound as the partial differential equation does in the $L_{2}$ norm, plus an added dissipation that depends on how much the approximate solution fails to satisfy the Rankine-Hugoniot jump.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the stability issue of the discontinuous Galerkin spectral element method (DGSEM) when solving hyperbolic equations with jump - coefficient matrices. Specifically, the author focuses on how to ensure the stability of the DGSEM method when the coefficient matrix jumps at the material interface.
### Problem Background
In wave - propagation problems, material properties (such as wave speed or density) will change abruptly at the junctions of different media. These interfaces will cause discontinuous jumps in the numerical solution, which brings difficulties to many numerical methods. For example, the interface between two dielectric materials in electromagnetic - wave - propagation problems; or the interface between different rock strata in geophysics.
### Research Motivation
For hyperbolic equations with jump coefficients, traditional energy - estimation methods (such as the L2 norm) may not directly give a stable conclusion. This is because even under homogeneous and dissipative boundary conditions, the L2 norm may increase or decrease, depending on the relative magnitudes of the wave speeds on both sides of the interface. Therefore, a new method needs to be found to evaluate the stability of such problems.
### Core Problem of the Paper
The core problem of the paper is: **How to use the L2 norm as an alternative indicator to infer the stability of hyperbolic equations with jump - coefficient matrices under the discontinuous Galerkin spectral element method (DGSEM)?**
### Solution
The author proposes a method based on the L2 norm to infer the stability of the DGSEM method. The specific steps are as follows:
1. **Energy Estimation**: First, analyze the energy behavior of the partial differential equation (PDE), and then obtain the energy behavior of the approximate solution through the discretization path.
2. **Rankine - Hugoniot Conditions**: Introduce the upwind numerical flux and ensure that it satisfies the Rankine - Hugoniot conditions (i.e., the conservation conditions). This can ensure that the discretized system has the same energy bound in the L2 norm and adds a dissipation term related to the deviation of the approximate solution from the Rankine - Hugoniot jump.
3. **Discounted Norm**: For some cases, although the L2 norm is not directly bounded, a discounted norm can be constructed so that the energy is bounded under this norm.
### Formula Summary
- L2 norm:
\[
||u||_{L^2}=\sqrt{\int_{\Omega}u^T u \, d\mathbf{x}}
\]
- Energy contribution at the interface:
\[
Q = -\frac{1}{2}(\mathbf{w}_L^T\bar{\Lambda}_L\mathbf{w}_L-\mathbf{w}_R^T\bar{\Lambda}_R\mathbf{w}_R)
\]
- Energy estimation in the discounted norm:
\[
\frac{d}{dt}||u||_{\alpha_c}^2\leq0
\]
where \(\alpha_c\) is the discount factor selected according to the wave - speed ratio.
### Conclusion
Through the above methods, the author proves that the stability of hyperbolic equations with jump - coefficient matrices under the DGSEM method can be effectively evaluated by the L2 norm. Although the L2 norm itself may not be directly bounded, by constructing an appropriate discounted norm, the energy of the system can be ensured to be bounded, thus guaranteeing the stability of the numerical method.
This method not only provides theoretical support for the discontinuous Galerkin spectral element method in dealing with complex physical problems, but also provides new ideas for the stability analysis of other numerical methods.