A high order, block finite difference, error inhibiting scheme for the transport equation

Adi Ditkowski,Anne Le Blanc,Chi-Wang Shu

DOI: https://doi.org/10.48550/arXiv.2402.11617

2024-02-18

Abstract:We propose a block finite difference, error inhibiting scheme that is fourth-order accurate for short to moderate times and has a six-order convergence rate for long times. This scheme outperforms the standard fourth-order Finite Difference scheme. We also demonstrate that the proposed scheme is a particular type of nodal-based Discontinuous Galerkin method with $p=1$.

Numerical Analysis

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the accuracy and stability of the numerical solution of the transport equation, especially in the case of long - time integration. Specifically:

1. **Reduce phase error**: In traditional methods, during long - time simulations, the overall error increases exponentially due to the linear growth of phase error. Therefore, the paper proposes a new numerical scheme to suppress this error accumulation.

2. **Improve the order of convergence**: The new scheme achieves fourth - order accuracy in short to medium time and has a sixth - order convergence rate in the long - time, which is far superior to the traditional fourth - order finite - difference method.

3. **Stability and high accuracy**: This scheme not only improves the accuracy but also ensures the stability of the numerical solution, avoiding the exponential growth of error that may occur under the classical Lax stability.

### Main contributions
- **Propose the Block Finite Difference (BFD) method**: Suppress error accumulation through the carefully designed interaction between the truncation error and the scheme dynamics.
- **Analyze the stability and accuracy of the BFD method**: Use the properties of eigenvalues and eigenvectors to prove the Von - Neumann stability of the scheme and show its convergence behavior under different conditions.
- **Show the relationship with the Discontinuous Galerkin (DG) method**: Prove that the proposed BFD method can be regarded as a node - based DG method, thus providing a theoretical basis for understanding its high - order accuracy.

### Mathematical formula representation
The mathematical formulas involved in the paper are represented in Markdown format as follows:

- Equation:
\[
\frac{\partial u}{\partial t}+\frac{\partial u}{\partial x} = 0,\quad x\in(0, 2\pi),\quad t\geq0
\]

- Truncation error:
\[
\tau_{j-\frac{1}{4}}^n=\frac{1}{11520}\left[120(p_1 + p_2)h^3u^{(4)}-(24 - 60p_2)h^4u^{(5)}\right]+O(h^5)
\]

- Eigenvalue expressions:
\[
\hat{\lambda}_1(k)=\frac{1}{12h}\left(\Omega-\sqrt{2\Delta}\right)
\]
\[
\hat{\lambda}_2(k)=\frac{1}{12h}\left(\Omega+\sqrt{2\Delta}\right)
\]
where:
\[
\Omega=-4i(p_1 + p_2 + 1)\sin(kh)+6(p_2 - p_1)\cos(kh)+10(p_2 - p_1)
\]
\[
\Delta=-80p_1 - 256+55p_2+55p_2^2-2(63p_1 + 40)p_2+4\left(15p_2+p_1(16 - 30p_2)+p_2(15p_2 + 16)+64\right)\cos(kh)+\left(13p_2+2p_1(8 - 5p_2)+p_2(13p_2 + 16)\right)\cos(2kh)+8i(-p_1 + p_2)\sin(kh)\left((3p_1 + 3p_2 + 4)\cos(kh)+5p_1 + 5p_2 + 28\right)
\]

Through these improvements, this paper provides a more efficient and accurate numerical method to solve the long - term integration problem in the transport equation.