Provably realizability-preserving finite volume method for quadrature-based moment models of kinetic equations

Chuan Fan, Qian Huang, Kailiang WuarXiv 2025 · DOI · arXiv

Introduction

Kinetic equations, such as the Boltzmann equation, provide a fundamental description of gas and plasma dynamics by tracking particle distributions in a high-dimensional phase space. While highly accurate, direct numerical solutions are computationally prohibitive for many engineering applications. Quadrature-Based Moment Methods (QBMM) offer a more efficient alternative by evolving a finite set of moments—such as density, momentum, and energy.

However, a persistent challenge in QBMM is maintaining "realizability": ensuring that the evolved moments always correspond to a physically valid underlying distribution. When numerical errors push moments outside the realizable set, simulations typically suffer from catastrophic failure, manifesting as negative temperatures or impossible pressure ratios. This paper introduces a robust finite volume framework designed to mathematically guarantee realizability even under extreme conditions.

Methodology: A Unified Framework

The authors develop a framework applicable to sophisticated models like Gaussian-EQMOM and HyQMOM. Their approach is built upon three primary pillars:

1. Geometric Quasi-Linearization (GQL)

The core innovation lies in the use of Geometric Quasi-Linearization to transform the complex, nonlinear inequalities defining the moment space into a manageable form. The authors prove that the realizability of moments can be characterized by the positive definiteness of specific Hankel matrices. By recasting these constraints as non-negative quadratic forms, the researchers provide a simpler mathematical foundation for ensuring stability through linear algebra tools.

2. Tailored HLL Numerical Flux

The framework employs a specifically constructed Harten–Lax–van Leer (HLL) numerical flux. By rigorously calculating wave speeds and intermediate states based on the underlying physics of the moments, the authors ensure that the flux itself "embeds" the realizability property. This ensures that the cell-averaged updates remain within the safe physical zone.

3. Realizability-Preserving Limiters

To achieve high-order accuracy without introducing non-physical oscillations, the method incorporates a specialized limiter. For second-order schemes, solvers must reconstruct values at the edges of grid cells. This limiter adaptively nudges these reconstructed values back into the realizable "safe zone" without sacrificing the overall accuracy of the simulation.

Numerical Validation

The proposed method was evaluated using a suite of classic, high-difficulty numerical benchmarks to verify its theoretical properties:

  • Smooth Accuracy Tests: These tests confirmed that the method maintains the expected second-order mathematical convergence rates.
  • Riemann and Shock Tube Problems: These benchmarks tested the solver's ability to handle sudden jumps in pressure and density, demonstrating that the scheme remains oscillation-free near shocks.
  • Shu–Osher Problem: This involved simulating the complex interaction between a shock wave and high-frequency density fluctuations, requiring both high accuracy and stability.
  • Double-Rarefaction Problem: An extreme test case involving near-vacuum regions. While standard QBMM solvers typically crash when density drops toward zero, this method successfully maintained positive weights and valid moments.

Key Insights and Results

The primary contribution of this work is the provable safety of the algorithm. The authors provide rigorous mathematical proofs that, under specific time-step (CFL) conditions, the scheme will never produce a non-realizable state. Furthermore, the framework is modular; the underlying physics model (e.g., EQMOM vs. HyQMOM) can be swapped by adjusting coefficients in the quadratic form.

Additionally, the scheme demonstrates multiscale stability. It remains robust regardless of the "relaxation time" (particle collision frequency), making it "asymptotic-preserving." This allows the solver to work effectively for both rarefied gases and dense, fluid-like gases.

Conclusion and Future Work

This research provides a mathematically rigorous solution to the stability issues that have long plagued quadrature-based moment methods. By ensuring realizability through GQL and tailored numerical fluxes, the authors have created a robust tool for complex fluid dynamics.

While currently demonstrated for one-dimensional problems and five-moment systems, the theoretical foundation suggests scalability. Future work will focus on implementing this framework in three-dimensional geometries and extending it to higher-order moment systems to capture even more complex physical phenomena.

Summary updated

BibTeX
@misc{fan2025provably,
  author = {Chuan Fan and Qian Huang and Kailiang Wu},
  title = {{Provably realizability-preserving finite volume method for quadrature-based moment models of kinetic equations}},
  year = {2025},
  eprint = {2510.18380},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2510.18380},
}
Figure