On algebraically stabilized schemes for convection-diffusion-reaction problems

Volker John,Petr Knobloch
DOI: https://doi.org/10.48550/arXiv.2111.08697
2021-11-17
Abstract:An abstract framework is developed that enables the analysis of algebraically stabilized discretizations in a unified way. This framework is applied to a discretization of this kind for convection-diffusion-reaction equations. The definition of this scheme contains a new limiter that improves a standard one in such a way that local and global discrete maximum principles are satisfied on arbitrary simplicial meshes.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is that in the numerical solution of convection - diffusion - reaction equations, most discretization methods cannot satisfy the discrete maximum principle (DMP), and thus there are problems in physical consistency. Specifically: 1. **Problem of physical consistency**: The solution of the convection - diffusion - reaction equation should satisfy the maximum principle, that is, the upper and lower bounds of the solution should be in line with the physical reality. However, many numerical methods cannot guarantee this property, especially in the convection - dominated case, the standard Galerkin finite - element method will produce inaccurate results, such as excessive smoothing in the layered region. 2. **Convection - dominated problem**: In the convection - dominated case ($\varepsilon \ll L\|b\|_{0,\infty,\Omega}$), the solution will have a steep gradient (called a layer) in some regions, and the thickness of these layers is usually much smaller than the grid resolution. Standard numerical methods perform poorly in this case, and a stabilized discretization method needs to be used to handle these problems. 3. **Limitations of existing methods**: Although the linearly stabilized finite - element method can satisfy the DMP, it usually has strict restrictions on the grid type and the results are not accurate enough. Although nonlinear methods can improve accuracy, they are difficult to solve in some cases, especially for the strongly convection - dominated case. To solve the above problems, this paper proposes a new algebraic stabilization scheme, aiming at: - Satisfying the local and global discrete maximum principle (DMP) on any triangular grid. - Maintaining the physical consistency of the solution, especially in the convection - dominated case. - Improving the accuracy of the numerical solution, especially in capturing the layered region. By introducing an improved Kuzmin limiter, this scheme can provide a more accurate numerical solution while maintaining stability and is applicable to a wider range of grid types.