High order compact fully-discrete scheme for hyperbolic conversation laws

Tong Zhou,Haitao Dong,Shucheng Pan

2024-07-31

Abstract:Based on the solution formula method, a series of one-step fully-discrete schemes, such as FWENO/Full-WENO has been proposed. Storing the by-products conservative variables at the half points (grid center) and using them as interpolation information to construct high-order schemes, we obtain a new class of one-step compact fully-discrete schemes. The new scheme can be associate with various non-oscillatory strategies. This paper takes state-of-the-art WENO-JS method as an example and proposes a family of compact fully-discrete WENO scheme. Detailed analysis is conducted on accuracy, errors, computational cost, efficiency and their connection with Hermite interpolation. Meanwhile, we design a new entropy flux linearization strategy for Euler equations to enhance its robustness, and also develop a multi-dimensional method for this compact fully-discrete framework. Due to the new scheme is one-step and utilizes stored by-products information for interpolation, it has a significant advantage in efficiency. For one-dimensional Euler equations, compared to the original FWENO, the computational cost only increases by 20-40%, while is approximately one-third of WENO+RK3. For two-dimensional case, a new special dimension-by-dimension strategy is applied. Although there is an additional computing cost, numerical experiments show that the new scheme only needs about 1/10 to 1/13 cost of that for WENO+RK3 when obtaining similar or even better resolution results, indicating that the new scheme is more efficient than semi-discrete schemes based on RK methods.

Numerical Analysis

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to improve the efficiency and accuracy of numerically solving hyperbolic conservation laws (such as the Euler equations), especially in long - time evolution problems. Specifically, the author proposes a new high - order compact fully - discrete scheme to overcome the limitations of existing methods in computational efficiency and accuracy.

### Main problems and challenges:
1. **Computational efficiency**: Traditional semi - discrete schemes rely on the Runge - Kutta (RK) method for time discretization, resulting in high computational costs, especially in multi - dimensional cases.
2. **Accuracy limitations**: High - order RK methods are difficult to guarantee the total variation diminishing (TVD) property, and as the order increases, the computational complexity rises significantly.
3. **Spatial and temporal coupled discretization**: Designing efficient space - time coupled discretization schemes is very challenging, especially for complex multi - dimensional problems.

### Solutions proposed in the paper:
1. **Fully - discrete framework based on the solution formula method**: By transforming the hyperbolic conservation law into the Hamilton - Jacobi (HJ) equation and using flux linearization to obtain a quasi - exact solution, an efficient fully - discrete scheme is constructed.
2. **Store and use half - point information**: Calculate and store not only the conservative variables at the grid nodes but also those at the half - points, which are used for interpolation to construct a higher - order compact scheme.
3. **Combine the WENO reconstruction strategy**: Apply the advanced WENO - JS reconstruction method to ensure non - oscillatory and high - order accuracy.
4. **Entropy - conditioned flux linearization**: Introduce entropy conditions to select the parameters of flux linearization, enhancing the robustness of the numerical scheme, especially when dealing with discontinuities.

### Main contributions:
- Propose a new class of high - order compact fully - discrete WENO schemes (CFWENO), which significantly improve computational efficiency while maintaining high accuracy.
- Verify the superior performance of the new scheme in one - and two - dimensional Euler equations through numerical experiments. In particular, with similar or better resolution, the computational cost is only one - third to one - tenth of that of traditional methods.

### Formula summary:
- Discrete form of the HJ equation:
\[
u_i^{n + 1}=u_i^n-\frac{\Delta t}{\Delta x}\left(f_{i + 1/2}^n - f_{i - 1/2}^n\right)
\]
- Numerical flux expression:
\[
f_{i + 1/2}^n=\frac{f(u_i^n)+f(u_{i + 1}^n)}{2}-\frac{\Delta t}{2\Delta x}\left(a_{i + 1/2}^n\left(u_{i + 1}^n - u_i^n\right)\right)
\]
- WENO reconstruction weights:
\[
\omega_k=\frac{\alpha_k}{\sum_{l = 0}^{r - 1}\alpha_l}, \quad \alpha_k=\frac{\gamma_k}{(\epsilon+\beta_k)^2}
\]

Through these improvements, the paper successfully provides an efficient and accurate numerical method suitable for solving complex problems in fluid mechanics and other fields.