Small Rarefaction, Large Consequences: Limits of Navier Stokes Turbulence Simulations

The paper demonstrates that subtle rarefaction effects in turbulent flows can cause "constitutive degeneracy" in Navier-Stokes simulations, leading to significant errors in predicting surface shear stress and heat flux during rocket plume impingement.

The Gist

The "Broken Compass" in the Storm: Why Our Best Math Fails in Space

Imagine you are trying to predict how a massive, swirling hurricane will hit a coastline. To do this, you use a very reliable weather model. This model works perfectly for big things: it predicts the wind speed, the direction of the storm, and where the rain will fall. For decades, scientists have assumed that as long as the storm is "big" and "thick" (like air on Earth), this math will always work.

But a new research paper by Tian and Wu suggests that in the extreme environments of space—like when a rocket lands on the Moon—our "weather model" (the Navier-Stokes equations) has a hidden flaw. It’s like having a compass that works perfectly in a calm field, but suddenly starts spinning wildly and giving wrong directions the moment it enters a specific kind of storm.

Here is the breakdown of what they discovered.

---

1. The Two Ways of Seeing the World

To understand the problem, you have to understand the two "languages" scientists use to describe fluids (like air or rocket exhaust):

• The "Smooth Liquid" View (Navier-Stokes): This treats air like a continuous, smooth syrup. It’s great for most things, but it assumes the air is "thick" enough that every molecule is constantly bumping into its neighbors.
• The "Lego Brick" View (Boltzmann): This treats air as a collection of individual, tiny "Lego bricks" (molecules) flying around. This is much more accurate for thin, ghostly gases in space, but it is incredibly difficult and "expensive" for computers to calculate.

2. The Discovery: "Constitutive Degeneracy"

Usually, scientists think that if the air is "thick" enough, the "Smooth Liquid" view is fine. They thought that even if there is turbulence (swirling chaos), the tiny individual molecules wouldn't matter much.

They were wrong.

The researchers looked at a rocket plume hitting the lunar surface. They found a phenomenon they called "constitutive degeneracy."

The Analogy: Imagine you are driving a car on a highway. The "Smooth Liquid" math is like looking at the road as one continuous surface. But in certain spots—like a sharp, twisting turn or a sudden patch of ice—the "smoothness" of the road disappears. In these specific spots, the car doesn't just slide; it behaves in ways the "smooth road" math simply cannot predict.

In the rocket exhaust, there are tiny "shear layers" (areas where the air is twisting violently). In these specific zones, the "Smooth Liquid" math predicts that the stress in the air should actually flip directions or drop to zero. But the "Lego Brick" reality shows that the stress doesn't disappear; it stays strong because the individual molecules are still pushing against each other.

3. Why This Matters: The "Small Error, Big Crash" Problem

Because the "Smooth Liquid" math fails in these tiny, localized twisting zones, it misses the "big picture" of how much heat and pressure are hitting the lunar lander.

The researchers found that the standard math underestimated the heat and the physical stress on the surface by as much as 50%.

The Analogy: It’s like a structural engineer designing a bridge. They use a formula that says, "The wind will push with 10 pounds of force." They build the bridge to handle 20 pounds. But because their formula failed to account for a tiny, specific type of swirling wind, the real force is actually 50 pounds. The bridge collapses, not because the wind was "huge," but because the math failed to see how the small swirls added up to a big punch.

The Bottom Line

When we send robots and humans to the Moon or Mars, we rely on math to tell us if our heat shields will melt or if our landing legs will snap. This paper warns us that turbulence can "unmask" the limits of our math. Even if the gas seems "thick" enough, the chaotic swirling can create tiny pockets where our standard equations break down, leading to massive errors in our most important engineering calculations.

Technical Summary

Technical Summary: Small Rarefaction, Large Consequences

Title: Small Rarefaction, Large Consequences: Limits of Navier–Stokes Turbulence Simulations
Authors: Songyan Tian and Lei Wu (Southern University of Science and Technology)
Date: February 10, 2026

---

1. Problem Statement

The paper addresses a fundamental question in fluid mechanics: Can molecular rarefaction (non-equilibrium) effects significantly influence macroscopic turbulent flows, even when the global Knudsen number ($Kn$) is small?

Traditionally, it has been assumed that in turbulent regimes, rarefaction effects are dynamically passive and only become relevant at the smallest Kolmogorov scales. While previous research suggested that deviations from the Navier–Stokes (NS) equations are confined to high wavenumbers or thermal fluctuations, these studies largely focused on homogeneous turbulence. This paper investigates whether inhomogeneous turbulence—specifically in high-speed, complex environments like rocket plume impingement on a lunar surface—can trigger localized breakdowns of the NS constitutive relations.

2. Methodology

The authors employ a dual-framework approach to compare the conventional continuum description with a high-fidelity kinetic description:

• Configuration: A 2D simulation of a Lunar Landing Module (LLM) subjected to two symmetric, high-Mach (Ma = 5.45) nitrogen rocket plumes. The flow is characterized by a low global $Kn$($6.87 \times 10^{-4}$) and a high Reynolds number ($Re = 1.18 \times 10^4$), placing it nominally in the continuum regime.
• Kinetic Framework (The Benchmark): They solve the Boltzmann equation using a modified Rykov kinetic model and the Discrete Velocity Method (DVM). To capture turbulence without a turbulence model, they perform Direct Numerical Simulations (DNS) of the Boltzmann equation by introducing synthetic velocity perturbations (100 superimposed eddies) to initiate unsteady jet dynamics.
• Multiscale Framework (The Comparison): They use a coarse-grained GSIS–SST framework. This couples the General Synthetic Iterative Scheme (GSIS)—which is asymptotic-preserving and captures rarefaction—with the Shear Stress Transport (SST) turbulence model. This allows them to see how the NS-based turbulence model performs against the kinetic "truth."

3. Key Contributions: "Constitutive Degeneracy"

The central theoretical contribution is the identification of constitutive degeneracy. The authors demonstrate that in regions of intense flow turning, expansion, or recirculation (common in turbulent shear layers), the linear NS constitutive relation (which relates stress to strain rate) can undergo a sign reversal or collapse.

In these specific zones, the predicted Newtonian shear stress ($\sigma_{NS}$) approaches zero because the components of the velocity gradient cancel each other out. However, the higher-order non-equilibrium stresses ($\sigma_{Boltzmann}$) derived from the kinetic theory do not vanish. Consequently, even though the rarefaction is "weak" (low $Kn$), the relative error becomes "order-one" because the denominator (the NS stress) has collapsed.

4. Results

• Underprediction of Engineering Quantities: The NS-SST model significantly underpredicts critical surface loads. Specifically, it underpredicts peak surface shear stress by 25–30% and surface heat flux by approximately 50% compared to the Boltzmann-based solutions.
• Spatial Discrepancy: The errors are most pronounced in the impingement and recirculation zones, where the flow is most inhomogeneous.
• Validation of Mechanism: The direct Boltzmann simulations (Fig. 2) confirm that the "degeneracy" is not an artifact of the coarse-grained SST model but a physical reality of the flow. The Boltzmann stress remains finite and becomes the dominant driver of momentum transport in regions where the NS stress crosses zero.
• Scale Interaction: The results show that turbulence and rarefaction are not independent; turbulence creates the complex, inhomogeneous velocity gradients that expose the limits of the NS equations.

5. Significance

This work has profound implications for aerospace engineering and planetary exploration:

1. Limits of CFD: It proves that standard Navier–Stokes-based CFD (Computational Fluid Dynamics) can be fundamentally unreliable for predicting thermal and structural loads in high-speed, turbulent, inhomogeneous flows, even if the gas appears to be a continuum.
2. Safety Margins: In missions involving rocket plume impingement (e.g., lunar or Mars landings), relying on NS equations may lead to a dangerous underestimation of material degradation and thermal stress.
3. Theoretical Insight: It resolves a long-standing debate by showing that rarefaction effects are not just a "small-scale" phenomenon but can be "macroscopically decisive" through the mechanism of constitutive degeneracy.