Consistent kinetic modeling of compressible flows with variable Prandtl numbers: Double-distribution quasi-equilibrium approach

This paper presents a consistent kinetic modeling and discretization strategy using a double-distribution quasi-equilibrium approach that enables accurate, stable, and Galilean-invariant simulations of compressible flows across all Prandtl numbers and specific heat ratios, successfully recovering Navier-Stokes-Fourier physics for both moderate supersonic speeds and complex discontinuities.

The Gist

Imagine you are trying to simulate how a fluid (like air or water) moves, but with a twist: you want to model it not just as a smooth flow, but as a swarm of billions of tiny, invisible particles bumping into each other. This is the world of kinetic modeling.

The paper you're asking about is like a master blueprint for building a super-accurate simulation engine that can handle these particles under extreme conditions—like when air is rushing around a supersonic jet or when heat is moving through a gas in weird ways.

Here is the breakdown of their breakthrough, explained with everyday analogies.

1. The Problem: The "One-Size-Fits-All" Trap

For a long time, scientists had a standard way to model these particles (called the BGK model). Think of this like a factory assembly line where every worker (particle) has the exact same job and speed.

• The Issue: Real life isn't like that. Sometimes, heat moves through a gas much faster or slower than momentum (how the gas pushes). This ratio is called the Prandtl number.
• The Old Way: The old models were like a factory that only worked if the heat and momentum moved at the exact same speed. If you tried to simulate a gas where heat moved differently (which happens in real life), the model broke or gave wrong answers. It was like trying to drive a car with the brakes and the accelerator permanently stuck together.

2. The Solution: The "Double-Track" System

The authors developed a new system using Double-Distribution Functions (DDF).

• The Analogy: Imagine a busy train station.
  • Track A (The f distribution): This track carries the "heavy stuff"—the mass and the momentum (the push). It's like the freight trains carrying the cargo.
  • Track B (The g distribution): This track carries the "energy stuff"—the heat and internal energy. It's like the passenger trains carrying the warmth.
• Why it helps: By separating the cargo from the passengers, the scientists can control how fast the cargo moves versus how fast the heat spreads. This allows them to simulate gases with any Prandtl number, whether heat moves fast, slow, or exactly in sync with the push.

3. The Secret Sauce: The "Quasi-Equilibrium" (QE)

Even with two tracks, you need a rulebook for how the particles behave when they crash or settle down. This is where the Quasi-Equilibrium (QE) approach comes in.

• The Analogy: Think of a chaotic dance floor.
  • Equilibrium: Everyone is dancing perfectly in sync (calm, stable).
  • Real Life: People are bumping into each other, spinning, and moving chaotically.
  • The QE Approach: Instead of forcing everyone to instantly snap back to perfect order, the model introduces an intermediate step. It says, "Okay, let's first get everyone to a 'semi-organized' state (Quasi-Equilibrium), and then let them settle into perfect order."
• The Benefit: This two-step relaxation process allows the model to handle the "slow" variables (like heat flow) and "fast" variables (like pressure) differently. It's like a traffic controller who lets the slow-moving trucks take a different route than the fast sports cars, preventing gridlock and ensuring everyone arrives safely.

4. The Construction: Building a Better Lattice

To make this work on a computer, they had to build a specific grid (a "lattice") to catch these particles.

• The Analogy: Imagine trying to catch raindrops with a net.
  • If the net has big holes (low resolution), you miss the small drops, and your simulation is inaccurate.
  • If the net is too small, the drops bounce off the edges.
• The Innovation: The authors used high-order velocity lattices. Think of this as a super-fine, multi-layered net with very specific hole sizes. They didn't just guess the hole sizes; they mathematically proved that this specific net catches exactly the right amount of momentum and energy to recreate the laws of physics (the Navier-Stokes equations) perfectly.

5. The Proof: Stress-Testing the Engine

The authors didn't just build the engine; they drove it through a hurricane to prove it works.

• Test 1: The Shock Tube (The Explosion): They simulated a sudden release of pressure (like a balloon popping). The model correctly predicted how the shockwave moved, regardless of the gas's heat properties.
• Test 2: The Thermal Couette Flow (The Hot Sandwich): They simulated two plates, one hot and moving, one cold and still, with gas in between. The model perfectly predicted the temperature profile, showing it understood how friction creates heat.
• Test 3: Shock-Vortex Interaction (The Whirlwind): This is the "final boss" test. They smashed a shockwave into a swirling vortex. This is incredibly sensitive; if your math is slightly off, the swirl looks wrong. Their model matched the most complex, high-precision computer simulations available, proving it captures the subtle physics of sound waves and heat dissipation.

The Big Picture

Why does this matter?

• Versatility: Before this, you needed different tools for different gases. Now, this single framework works for any gas, at any temperature ratio, and with any heat-to-momentum ratio.
• Efficiency: It's fast and stable, meaning engineers can use it to design better jet engines, predict weather patterns, or study how heat moves in space vehicles without the simulation crashing.
• The Future: This paper lays the groundwork for simulating even more extreme scenarios, like hypersonic flight (faster than Mach 5) or the complex interactions inside fusion reactors.

In short: The authors built a universal, high-precision simulation engine that separates "push" from "heat," uses a clever two-step relaxation rule to keep things stable, and has been proven to work perfectly even in the most chaotic, high-speed fluid scenarios.

Technical Summary

1. Problem Statement

The simulation of compressible fluid flows, particularly those involving high Mach numbers, variable specific heat ratios ($\gamma$), and non-unity Prandtl numbers ($Pr$), presents significant challenges for numerical methods.

• Limitations of Standard Models: The standard Bhatnagar–Gross–Krook (BGK) model restricts the Prandtl number to unity ($Pr=1$) and assumes monoatomic gases (fixed $\gamma = 1.4$ or $5/3$).
• Existing Gaps: While Double-Distribution Function (DDF) approaches have been developed to handle variable $\gamma$ (by introducing a second distribution function $g$ for internal energy), and Quasi-Equilibrium (QE) approaches exist to handle variable $Pr$, a unified, consistent framework combining both DDF and QE for all Prandtl numbers ($Pr \in [0, \infty)$) and specific heat ratios was lacking.
• Consistency Issues: Previous attempts often failed to strictly recover all macroscopic moments and dissipation rates required by the Navier-Stokes-Fourier (NSF) equations, or lacked strict conservation of mass, momentum, and total energy across all regimes.

2. Methodology

The authors developed a consistent kinetic modeling and discretization strategy using the Quasi-Equilibrium (QE) approach within two distinct Double-Distribution Function (DDF) frameworks.

A. Kinetic Model Formulation

• Two-Step Relaxation: The model employs a two-step relaxation process (QE-BGK) to separate fast and slow modes.
  • Fast relaxation ($\tau_1$): From the current state $f$ to an intermediate quasi-equilibrium state $f^*$.
  • Slow relaxation ($\tau_2$): From $f^*$ to the local Maxwell-Boltzmann equilibrium $f^{eq}$.
  • This hierarchy ($\tau_1 \leq \tau_2$) ensures the H-theorem (entropy production $\leq 0$) and allows independent control of shear viscosity ($\mu$) and thermal conductivity ($\kappa$), thereby controlling $Pr$.
• Double-Distribution Strategy:
  • Distribution $f$: Carries mass and momentum.
  • Distribution $g$: Carries energy. Two specific energy splitting strategies were analyzed:
    1. Total Energy Split: $g$ carries the total energy.
    2. Internal Non-Translational Energy Split: $g$ carries only the internal energy associated with non-translational degrees of freedom (roto-vibrational modes).
• Quasi-Equilibrium Constraints: To recover variable $Pr$, the quasi-equilibrium states ($f^*, g^*$) are constructed by minimizing the H-function subject to specific constraints:
  • For $Pr \leq 1$: The pressure tensor is set to equilibrium, while the heat flux is set to the full off-equilibrium distribution.
  • For $Pr \geq 1$: The pressure tensor is set to the off-equilibrium distribution, while the heat flux is set to equilibrium.
  • This distinction ensures the correct recovery of the NSF equations for any $Pr$.

B. Discretization Strategy

• Discrete Velocity Boltzmann (DVB): The model is discretized in phase space using high-order velocity lattices (specifically D2Q16 and D2Q25 in 2D) based on Gauss-Hermite quadrature.
• Static Reference Frame: Unlike some hypersonic methods that use moving lattices, this approach uses a static reference frame but employs high-order expansions (Grad-Hermite) to accurately capture higher-order moments without correction terms.
• Moment Recovery:
  • Total Energy Split: Requires $f$ to recover moments up to order 3 and $g$ up to order 2.
  • Internal Energy Split: Requires $f$ to recover moments up to order 4 and $g$ up to order 2.
• Numerical Scheme: A time-explicit finite-volume scheme with a first-order Euler forward time discretization and a Nearest Neighbor Deformation (NND) interpolation scheme with a generalized van Leer limiter for flux reconstruction.

3. Key Contributions

1. Unified Framework: The first consistent kinetic model combining DDF and QE approaches valid for all Prandtl numbers ($Pr \in [0, \infty)$) and specific heat ratios.
2. Rigorous Hydrodynamic Limit: A detailed Chapman-Enskog analysis proves that the model recovers the full Navier-Stokes-Fourier equations, including correct expressions for:
   • Shear viscosity ($\mu$) and bulk viscosity ($\eta$).
   • Thermal conductivity ($\kappa$).
   • Viscous heating terms.
3. Strict Conservation: The discretization ensures strict conservation of mass, momentum, and total energy at the machine precision level.
4. Two Energy Split Variants: The paper provides a comparative analysis of two energy splitting methods (Total vs. Internal Non-Translational), detailing their respective requirements for velocity set order and stability limits.

4. Results and Validation

The models were validated against a series of benchmarks demonstrating accuracy, stability, and Galilean invariance:

• Conservation Properties: A Sod shock tube test confirmed that all models (for $Pr=0.5, 1, 2$) are strictly conservative, with errors at the level of machine epsilon ($\approx 10^{-16}$).
• Dispersion and Dissipation:
  • Speed of Sound: Correctly recovered across a wide temperature range (4 orders of magnitude for Total Split, 3 for Internal Split) and Mach numbers up to $Ma \approx 3.0$.
  • Viscosity and Diffusivity: Shear viscosity, bulk viscosity, and thermal diffusivity were measured and found to match analytical predictions derived from the Chapman-Enskog expansion for various $Pr$ and $Ma$.
• Thermal Couette Flow: Simulations of viscous heating and thermal conduction at $Ma=0.5, 1.2$ and $Ec=4, 40$ showed excellent agreement with analytical solutions for various Prandtl numbers.
• Shock-Vortex Interaction: A sensitive benchmark involving a shock wave interacting with a vortex ($Ma_s \approx 1.2-1.29$). The models accurately reproduced:
  • The deformation of the shock.
  • The generation of sound waves (acoustic pressure).
  • Quantitative sound pressure distributions matching Direct Numerical Simulation (DNS) references and experimental data.

5. Significance and Outlook

• Scientific Impact: This work establishes a robust, thermodynamically consistent framework for simulating complex compressible flows where thermal and momentum diffusion rates differ significantly (non-unity $Pr$), which is common in real-world applications (e.g., combustion, atmospheric re-entry, polyatomic gas dynamics).
• Practical Utility: The models offer high accuracy and numerical stability without the need for ad-hoc correction terms, making them suitable for studying complex fluid dynamics problems.
• Future Directions: The authors note that while the current work uses high-order static lattices, future extensions will focus on:
  • Applying correction terms to allow the use of standard, lower-order lattices (e.g., D2Q9).
  • Extending the method to hypersonic regimes using target-designed reference frames (e.g., the "Particles on Demand" method).
  • Implementing space-time adaptive refinement to improve computational efficiency.

In summary, the paper provides a mathematically rigorous and numerically validated solution to a long-standing challenge in kinetic theory: simulating compressible flows with arbitrary thermodynamic properties and transport coefficients within a unified double-distribution framework.