TL;DR
This paper introduces novel third-order accurate Discontinuous Galerkin (DG) schemes for solving convection-diffusion problems with anisotropic diffusivity. These schemes are designed to rigorously preserve the maximum principle, ensuring numerical solutions remain within physically meaningful bounds, a crucial feature for many real-world applications.
What Problem Does This Paper Address?
Convection-diffusion equations are fundamental in modeling various physical phenomena, from heat transfer to fluid dynamics. A key property of solutions to many such problems is the "maximum principle," which guarantees that the solution remains bounded within the range defined by the initial data. This is critical for physical realism and numerical stability. However, developing high-order accurate numerical methods that also strictly preserve this maximum principle, especially in the presence of anisotropic diffusion (where heat or mass spreads at different rates in different directions), is a significant challenge. Existing methods often struggle to maintain both high accuracy and the maximum principle simultaneously, leading to unphysical oscillations or even blow-ups.
Methods
The authors employ the Discontinuous Galerkin (DG) method, a powerful technique for solving partial differential equations. Specifically, they build upon the Direct Discontinuous Galerkin (DDG) framework, which has shown promise for diffusion problems. The core of their contribution lies in:
- Third-Order Accurate DG Scheme: They construct a DG scheme that achieves third-order accuracy in space.
- Anisotropic Diffusion Handling: The scheme is designed to handle anisotropic diffusion tensors, which are more complex than isotropic diffusion.
- Maximum Principle Preservation: The key innovation is the integration of a "scaling limiter" (previously proposed by the authors) with the DDG scheme. This limiter, applied locally, ensures that the numerical solution's cell averages stay within the bounds dictated by the initial data, without sacrificing the scheme's accuracy.
- Weighted Cell Averages: The method utilizes weighted cell averages, which are crucial for both ensuring weak monotonicity and applying the scaling limiter effectively, particularly in the context of anisotropic diffusion.
- Time Discretization: For time integration, they use Strong Stability Preserving (SSP) Runge-Kutta methods, which are known to be compatible with high-order spatial discretizations and help maintain stability.
Datasets
Not discussed in the paper. The paper focuses on developing and analyzing numerical methods for a class of PDEs. Numerical experiments are conducted using synthetic problems with known exact solutions or reference solutions obtained through refined discretizations.
Key Insights and Ideas
- The Power of the Scaling Limiter: The paper demonstrates that a carefully designed scaling limiter can effectively enforce the maximum principle for high-order DG schemes, even in the challenging case of anisotropic diffusion, without compromising accuracy.
- Weighted Cell Averages as a Unifying Tool: The use of weighted cell averages proves to be a versatile concept, instrumental in both the theoretical analysis of the maximum principle and the practical implementation of the limiter.
- Anisotropy's Impact on CFL Conditions: The presence of anisotropic diffusivity leads to more restrictive conditions on the time step (CFL conditions) required to maintain the maximum principle, highlighting a trade-off between accuracy and computational efficiency.
- Bridging Theory and Practice: The work successfully bridges theoretical analysis of maximum-principle preservation with practical numerical implementation and validation through extensive tests.
Key Results
- Uniform Third-Order Accuracy: The proposed DG schemes achieve uniform third-order accuracy for smooth solutions.
- Strict Maximum Principle Satisfaction: Numerical tests confirm that the schemes, when equipped with the scaling limiter, strictly satisfy the maximum principle, keeping solutions within the initial bounds [C1, C2] for all tested scenarios.
- Effective Handling of Anisotropy: The schemes demonstrate robustness in handling both isotropic and anisotropic diffusion tensors, including cases with highly anisotropic diffusion.
- Comparison with Un-limited Schemes: Numerical examples show that schemes without the MPS limiter can exhibit significant overshoots and oscillations, leading to blow-up in short times, underscoring the necessity of the proposed approach.
Limitations and Future Work
- CFL Condition Restrictiveness: The paper notes that the CFL conditions for preserving the maximum principle can be restrictive, especially in the two-dimensional case, potentially leading to slow numerical time evolution.
- Implicit Schemes: The authors suggest that extending the present results to implicit schemes could improve computational efficiency, particularly by mitigating the restrictive CFL conditions.
- Extension to Three Dimensions: While the methodology is stated to extend to three dimensions, detailed analysis and numerical results for 3D are not presented in this paper.
- Boundary Conditions: The analysis primarily focuses on periodic boundary conditions. Modifications for other types of boundary conditions are mentioned as a possibility but not elaborated upon.
Summary updated