Passive scalar cascade in the intermediate layer of turbulent channel flow for $Pr\leq 1$

This study investigates the similarities and differences between velocity and passive scalar scale-by-scale equilibria in turbulent channel flow with $Pr \leq 1$, revealing that while both fields achieve Kolmogorov equilibrium at a specific length scale $r_{min}$ below the inertial range, the characteristics of this equilibrium and the nature of their interscale transfer rates exhibit distinct Prandtl number dependencies and alignment behaviors.

The Gist

Imagine you are stirring a pot of soup. You have the big, swirling currents of the spoon (the velocity of the fluid) and you have the heat or spices spreading through the liquid (the passive scalar, like temperature or dye).

For a long time, scientists thought that the way heat spreads through a turbulent soup was just a simple copy of how the soup itself moves. But this paper says: "Not quite! There are some fascinating differences, especially when the heat spreads faster than the soup moves."

Here is a simple breakdown of what the researchers discovered, using everyday analogies.

1. The Setting: The "Middle Layer" of the Soup

The researchers looked at a specific part of the flow called the intermediate layer.

• The Analogy: Imagine a river. Near the banks, the water is slow and sticky. In the very center, it's fast and chaotic. The "intermediate layer" is the middle zone where the water is moving fast enough to be turbulent, but not so close to the center that the rules change completely.
• The Goal: They wanted to see if the "heat" (scalar) follows the same rules as the "water movement" (velocity) in this middle zone.

2. The Big Discovery: The "Sweet Spot" of Chaos

In turbulence, energy doesn't just disappear; it gets passed down from big swirls to tiny swirls, like a game of "hot potato" where the potato gets smaller and smaller until it vanishes as heat. This is called a cascade.

• The Old Idea: Scientists used to think this "hot potato" game happened smoothly across a wide range of sizes (the "inertial range").
• The New Finding: The paper shows that for both the water and the heat, the "perfect balance" (where the energy passing down equals the energy disappearing) only happens at one specific size.
• The Metaphor: Think of a waterfall. The water doesn't flow perfectly evenly all the way down. There is one specific ledge where the water hits the rock and splashes perfectly. The researchers found that for both the water and the heat, this "perfect splash" only happens at a very specific, tiny scale.
  • For the water, this scale is called the Taylor length.
  • For the heat, this scale is called the Batchelor length.

3. The Twist: Heat Moves Faster (The Prandtl Number)

The paper focuses on cases where the heat spreads faster than the water moves (like hot air or oil with low viscosity). This is measured by something called the Prandtl number ($Pr$).

• The Analogy: Imagine two runners. Runner A (Water) is heavy and slow. Runner B (Heat) is light and fast.
• The Result: Because Runner B is faster, the "perfect splash" (the equilibrium) happens at a different size than for Runner A.
  • The researchers found a mathematical rule: The size of this "sweet spot" for heat shrinks as the heat gets faster. Specifically, it shrinks by the cube root of the speed difference.
  • Simple takeaway: If you make the fluid "thinner" to heat (lower Prandtl number), the zone where heat balances out gets smaller and smaller.

4. The Hidden Difference: Aligned vs. Anti-Aligned

This is the most surprising part. While the overall behavior of the water and heat looks similar, the details of how they interact are different.

• The Analogy: Imagine a dance floor.
  • Aligned Pairs: Dancers moving in the same direction.
  • Anti-Aligned Pairs: Dancers moving in opposite directions (crashing into each other).
• The Finding:
  • For the water, the "crashing" dancers (anti-aligned) and the "moving together" dancers (aligned) both contribute significantly to the energy transfer.
  • For the heat, the "moving together" dancers contribute much less. The heat transfer is almost entirely driven by the "crashing" dancers.
• Why it matters: Even though the big picture looks the same, the microscopic "dance steps" the heat takes are different from the water. The heat is more sensitive to the chaotic collisions than the water is.

5. How They Did It

The researchers didn't just guess; they used two tools:

1. Mathematical Magic (Asymptotics): They used advanced math to predict what should happen if the turbulence was perfect and infinite.
2. Supercomputer Simulations (DNS): They created a virtual channel flow on a supercomputer, simulating millions of tiny particles of water and heat to see if the math held up.

The Bottom Line

This paper tells us that while heat and water flow look like twins in a turbulent river, they are actually cousins.

1. They both find a "sweet spot" where the energy transfer balances out, but that spot is at a different size depending on how fast the heat moves.
2. They both rely on chaotic collisions to move energy, but heat relies on these collisions much more heavily than water does.

Why should you care?
Understanding these tiny details helps engineers design better engines, predict how pollution spreads in the atmosphere, and improve how we mix chemicals in industrial processes. It turns out that even in a chaotic soup, there is a very specific, hidden order to how heat behaves.

Technical Summary

Here is a detailed technical summary of the paper "Passive scalar cascade in the intermediate layer of turbulent channel flow for $Pr \le 1$" by Gallorini et al.

1. Problem Statement

The study investigates the scale-by-scale energy budget and cascade mechanisms of passive scalar fields (specifically temperature) in the intermediate layer of a fully developed turbulent channel flow (TCF). While the behavior of turbulent kinetic energy (velocity fields) in this region has been recently analyzed (e.g., by Apostolidis et al., 2023), the corresponding behavior for passive scalars with Prandtl numbers $Pr \le 1$ (where scalar diffusivity is higher than momentum diffusivity) remains less understood.

Key questions addressed include:

• Does a Kolmogorov-type scale-by-scale equilibrium exist for passive scalars in non-homogeneous turbulence?
• If so, at what specific length scale does this equilibrium occur?
• How do the interscale transfer properties of the scalar field compare to those of the velocity field, particularly regarding aligned and anti-aligned velocity fluctuations?

2. Methodology

The authors employ a dual approach combining Direct Numerical Simulations (DNS) and Matched Asymptotic Expansions.

• DNS Setup:
  • Flow: Pressure-driven turbulent channel flow between isothermal walls with a uniform internal heat source to maintain a constant bulk mean temperature.
  • Parameters: Bulk Reynolds number $Re_b = 40,000$ (friction Reynolds number $Re_\tau = 1000$).
  • Prandtl Numbers: Simulations were conducted for $Pr = 1, 0.75, 0.5,$ and $0.25$.
  • Resolution: High-resolution spectral code with $512 \times 512$ Fourier modes in streamwise/spanwise directions and 515 points in the wall-normal direction.
• Theoretical Framework:
  • The study extends the Karman-Howarth-Monin-Hill (KHMH) equation approach (used for velocity) to the generalized Yaglom equation for passive scalars.
  • Matched Asymptotics: The authors apply the method of matched asymptotic expansions (Van Dyke, 1964) to the Yaglom equation. They hypothesize "homogeneous two-point physics" within the non-homogeneous turbulence of the intermediate layer.
  • Scaling: They derive outer (large scale, $y$) and inner (small scale, Kolmogorov/Batchelor) scalings for the second-order structure functions of velocity and scalar variance.

3. Key Contributions

• Extension of Equilibrium Theory: The paper demonstrates that, similar to velocity fields, a Kolmogorov-like equilibrium between interscale transfer and dissipation for passive scalars is not achieved across the entire inertial range. Instead, it is achieved asymptotically only around a specific length scale, $r_{min}$.
• Identification of the Critical Length Scale ($r_{min}$): The authors derive that the scale where the scalar cascade is most efficient (maximum transfer rate relative to dissipation) is the Taylor-Prandtl length scale:
  $$r_{min} \sim \lambda_T = \frac{\lambda}{Pr^{1/3}}$$
  where $\lambda$ is the Taylor microscale. This contrasts with the velocity field where $r_{min} \sim \lambda$.
• Derivation of Scaling Laws: The study provides power-law dependencies for $r_{min}$ and the minimum transfer-to-dissipation ratio based on the Prandtl number, derived from matched asymptotics without relying solely on dimensional analysis.
• Decomposition of Cascade Mechanisms: The paper introduces a detailed analysis of the interscale transfer rate decomposed into contributions from aligned and anti-aligned fluctuating velocity pairs, revealing distinct differences between scalar and velocity cascades.

4. Key Results

• Asymptotic Equilibrium: In the intermediate layer ($\delta_\nu \ll y \ll h$), the scalar interscale transfer rate ($\Pi^v_T$) balances the scalar dissipation rate ($\varepsilon^v_T$) only asymptotically around $r_{min} \approx \lambda_T$.
  • At scales larger than $r_{min}$, production terms dominate, causing a systematic departure from equilibrium.
  • At scales smaller than $r_{min}$, molecular diffusion dominates.
• Prandtl Number Dependence:
  • The location of the minimum transfer rate scales as $r_{min} \propto Pr^{-1/3}$.
  • The deviation from equilibrium scales as $1 + (\Pi^v_T/\varepsilon^v_T){min} \propto Pr^{-2/9} Re\lambda^{-2/3}$.
  • DNS data confirms these scaling laws, showing excellent collapse when variables are normalized by $Pr$.
• Scalar vs. Velocity Cascade:
  • Global Similarity: Both velocity and scalar fields exhibit a negative interscale transfer rate (forward cascade) on average, with minima occurring at similar normalized scales ($r/\lambda$ or $r/\lambda_T$).
  • Critical Differences: When decomposing the transfer rate into contributions from aligned ($\Rightarrow$) and anti-aligned ($\Leftrightarrow$) velocity pairs:
    • The anti-aligned contributions dominate the cascade for both fields and show similar behavior.
    • The aligned contributions differ significantly: for the scalar field, the magnitude of the aligned contribution is much lower than for the velocity field.
    • Crucially, for the velocity field, the peak of the anti-aligned contribution coincides with the zero-crossing of the aligned contribution. For the scalar field, this coincidence does not occur, suggesting that the scalar field does not tend toward a pure Kolmogorov limit ($\Pi/\varepsilon \to -1$) in the same way the velocity field might as $Re_\tau \to \infty$.

5. Significance

This work fundamentally advances the understanding of passive scalar transport in wall-bounded turbulence. By proving that the "equilibrium" scale for scalars is distinct from that of velocity (shifting from $\lambda$ to $\lambda/Pr^{1/3}$), it refines turbulence modeling for heat and mass transfer, particularly in low-Prandtl number regimes (e.g., liquid metals).

The finding that the scalar cascade does not fully decouple from production effects in the same manner as the velocity field, and the identification of discrepancies in aligned/anti-aligned contributions, suggests that scalar mixing is more sensitive to the specific flow structures (e.g., ejections and sweeps) than previously assumed. This has implications for improving subgrid-scale models in Large Eddy Simulations (LES) and understanding mixing efficiency in combustion and environmental flows.