TL;DR
This paper introduces a novel micro-macro decomposition method for simulating complex kinetic equations, like the Boltzmann and Fokker-Planck-Landau equations. The key innovation is a scheme that efficiently handles the stiffness of collision terms while exactly conserving fundamental physical quantities like mass, momentum, and energy, making it suitable for a wide range of physical regimes.
What Problem Does This Paper Address?
Simulating kinetic equations, which describe the behavior of many-particle systems, is notoriously challenging due to the presence of multiple time and spatial scales. This is often characterized by a small Knudsen number, representing the ratio of mean free path to a characteristic length scale. Explicit numerical methods struggle with the "stiffness" introduced by collision terms, requiring prohibitively small time steps. Implicit methods, while allowing larger time steps, are computationally expensive. Furthermore, ensuring that fundamental physical quantities (mass, momentum, energy) are conserved numerically is crucial for the accuracy and stability of simulations, especially in fluid-like regimes. Existing methods often fail to guarantee this exact conservation, requiring extra effort.
Methods
The authors extend the "micro-macro decomposition" (MM) method, previously applied to simpler kinetic models, to handle general collisional kinetic equations like the Boltzmann and Fokker-Planck-Landau equations. The core idea is to decompose the distribution function $f$ into a macroscopic part $M$ (related to equilibrium) and a microscopic part $g$ ($f = M + g$). This decomposition, combined with an implicit-explicit (IMEX) time discretization, leads to an Asymptotic-Preserving (AP) scheme.
A crucial aspect is addressing the stiffness of the linearized collision operator. They achieve this by using a penalization method. For the Boltzmann equation, they employ a BGK-based penalty, and for the Fokker-Planck-Landau equation, a linear Fokker-Planck penalty. This allows for efficient inversion of the stiff terms.
A significant contribution is the exact conservation of moments. This is achieved by simultaneously solving the kinetic equation for $g$ and a separate system of macroscopic equations for the moments ($U = (\rho, \rho u, E)^T$). By discretizing the macroscopic system using standard conservative schemes, the numerical conservation of mass, momentum, and energy is guaranteed.
Datasets
Not discussed in the paper. The paper focuses on developing and testing numerical methods for established kinetic equations.
Key Insights and Ideas
- Extending MM to General Collisions: The micro-macro decomposition, previously limited to simpler collision models, is successfully generalized to handle the complex Boltzmann and Fokker-Planck-Landau collision operators.
- Implicit-Explicit (IMEX) and Penalization for Stiffness: Combining IMEX time stepping with a penalization strategy for the collision operator provides an efficient way to handle stiffness, allowing for larger time steps without sacrificing accuracy.
- Exact Moment Conservation via Macroscopic System: The most striking insight is that by explicitly solving a system of macroscopic equations for the conserved moments, exact numerical conservation of mass, momentum, and energy is achieved, regardless of the discretization of the kinetic equation itself. This is a significant advantage over methods that derive moments directly from the discretized kinetic distribution function.
- Generic Approach: The strategy of solving both the kinetic and moment systems simultaneously is presented as a generic recipe applicable to a wide range of kinetic equations, including Vlasov-Poisson and Vlasov-Ampère systems.
Key Results
The paper demonstrates through numerical examples that the proposed micro-macro decomposition based AP schemes:
- Work in Multiple Regimes: They accurately capture fluid-like behavior even when the Knudsen number is small, without requiring extremely fine meshes or time steps.
- Conserve Moments Exactly: Numerical simulations show perfect conservation of mass, momentum, and total energy when these quantities are obtained from the solved macroscopic moment system. This is verified across various test cases, including those with varying Knudsen numbers and shock tube problems.
- Are Consistent with Existing Methods: The results from the proposed MM schemes are shown to be consistent with established methods like Filbet-Jin (FJ) and Jin-Yan (JY) for similar problems.
- Handle Vlasov-Maxwell Systems: The approach is extended to Vlasov-Ampère and Vlasov-Ampère-Boltzmann systems, demonstrating exact total energy conservation for these plasma physics models.
Limitations and Future Work
- First-Order Time Discretization: The current implementation uses a first-order IMEX temporal discretization. The authors suggest improving this to a second-order scheme for better accuracy.
- Multi-Dimensional Extension: Extending the micro-macro method to higher spatial dimensions is identified as a significant area for future research.
- Computational Cost: While efficient, solving both the kinetic and macroscopic systems simultaneously might incur higher computational costs compared to methods that only solve the kinetic equation.
- Numerical Total Energy Error: For Vlasov-Poisson-Landau systems, the numerical total energy exhibits an $O(10^{-3})$ error that persists over long times. The authors suggest investigating ways to diminish this error, possibly through conservation constraints in the kinetic simulation or improvements in operator splitting techniques.
Summary updated