Viscous spectral energy coupling across scales in generalised Newtonian fluids

This study demonstrates that in generalised Newtonian fluids, the nonlinear viscous term in the momentum equation acts not only as a dissipation mechanism but also as a conservative energy transfer agent that drives a forward cascade and replaces the classical exponential spectral cutoff with a power-law decay, particularly in shear-thickening regimes.

The Gist

Imagine a busy highway where cars represent tiny swirls of energy in a fluid (like water or air). In a normal, "Newtonian" fluid (like water), the rules of the road are simple:

1. The Convective Term (The Drivers): Drivers naturally change lanes and interact with neighbors. This is how energy moves from big, slow-moving trucks (large scales) to tiny, fast-moving motorcycles (small scales). This is the only way energy usually travels down the highway.
2. The Viscous Term (The Friction): Friction acts like a brake. It slows down the cars and turns their speed into heat. In normal fluids, this brake is constant and acts locally—it just stops the car right where it is, without moving energy to other cars.

The Big Discovery
This paper investigates what happens when the "road conditions" change. Imagine a fluid where the "friction" (viscosity) isn't constant. Instead, it changes depending on how fast the cars are moving or how crowded the road is. This is called a "generalized Newtonian fluid."

The researchers used powerful computer simulations to watch these fluids behave. They found something surprising: When the friction changes, the "brakes" start acting like "drivers."

Here is the breakdown of their findings using everyday analogies:

1. The "Brake" Becomes a "Traffic Cop"

In a normal fluid, the friction term is just a simple brake. But in these special fluids, because the friction changes from spot to spot, the math shows that the friction term becomes nonlinear.

Think of it this way: In a normal fluid, the brake just slows you down. In these special fluids, the brake system is so complex that it starts shuffling energy between different cars. It doesn't just stop a car; it takes energy from a slow truck and gives it to a fast motorcycle, or vice versa.

The paper proves that this "viscous shuffling" is real. It behaves mathematically just like the drivers shuffling energy, even though it comes from the friction term.

2. Two Different Fluids, Two Different Stories

The researchers tested two types of these special fluids, and they behaved very differently:

• Shear-Thinning Fluids (The "Runaway" Fluid):

  • Analogy: Imagine a fluid that gets thinner and slipperier when you stir it fast (like ketchup or paint).
  • Result: When the fluid gets thin in high-speed areas, the "brakes" actually start acting like a gas pedal. They add a little bit of energy back into the system in those specific spots. However, they don't really shuffle energy between different sizes of swirls. The energy still moves down the highway mostly via the "drivers" (convection), and the tiny swirls die out very quickly (exponentially), just like in normal water.

• Shear-Thickening Fluids (The "Jamming" Fluid):

  • Analogy: Imagine a fluid that gets thicker and stiffer when you stir it fast (like a mixture of cornstarch and water, or "Oobleck").
  • Result: This is where the magic happens. When the fluid gets stiff in high-speed areas, the "brakes" transform into a super-efficient traffic cop.
  • They found a specific pattern (a "dipole") where the friction actively takes energy from one size of swirl and passes it to a slightly smaller one.
  • The Consequence: Because this "friction traffic cop" is helping move energy down the line, the tiny swirls don't die out as fast as they usually do. Instead of disappearing instantly (exponential decay), they linger and follow a predictable, slower pattern (power-law decay). It's as if the friction is keeping the tiny motorcycles running longer than physics usually allows.

3. The "Traffic Jam" at the End of the Highway

In normal fluids, once energy reaches the smallest scales, it vanishes instantly into heat. The graph of energy drops off a cliff.

In the "shear-thickening" fluids studied, because the friction is helping to pass the energy along, the energy doesn't drop off a cliff. Instead, it slides down a gentle ramp. The paper shows that this "ramp" (power-law decay) is a direct result of the friction term taking over the job of moving energy when the fluid gets very small and stiff.

4. Why This Matters (According to the Paper)

The paper makes a fundamental point about how we understand physics:

• Old Belief: Only the "drivers" (convection) can move energy between different sizes of swirls. The "brakes" (viscosity) only stop things.
• New Reality: Any part of the equation that gets complicated (nonlinear) can start moving energy. If the friction changes based on the flow, the friction itself becomes a mechanism for moving energy across scales.

The authors also note a connection to Large Eddy Simulation (LES), a method used by engineers to simulate complex flows. Many of these simulations use a "fake friction" (eddy viscosity) that acts exactly like the "shear-thickening" fluid in this study. The paper predicts that if you look closely at the data from these simulations, you should see this same "friction traffic cop" behavior and the resulting "gentle ramp" of energy decay, because the math is identical.

Summary

In short, this paper shows that in fluids where the "stickiness" changes with speed, the friction doesn't just stop the flow—it starts helping to shuffle energy around. In fluids that get stiffer when stirred (shear-thickening), this friction becomes so effective at shuffling energy that it changes the very way the fluid's smallest swirls disappear, turning a sudden stop into a gradual slide.

Technical Summary

Technical Summary: Viscous Spectral Energy Coupling in Generalised Newtonian Fluids

Problem Statement
In classical Newtonian turbulence, the spectral energy budget is governed by a strict separation of roles: the nonlinear convective term redistributes energy across scales via conservative triadic interactions (the energy cascade), while the linear viscous term acts as a purely local, non-conservative dissipation mechanism proportional to $\nu\kappa^2\hat{E}(\kappa)$. However, in generalised Newtonian fluids, viscosity is not constant but a spatially and temporally varying field coupled to the local strain rate. This introduces an additional nonlinear term in the Navier-Stokes equations. The spectral structure and physical interpretation of this variable viscosity term have received little attention. Specifically, it is unclear whether the resulting viscous mode-to-mode coupling carries conservative effects (energy transfer) or remains purely dissipative, and how this coupling influences the spectral energy budget in both shear-thinning and shear-thickening regimes.

Methodology
The authors investigate these dynamics using Direct Numerical Simulation (DNS) of homogeneous isotropic turbulence for generalised Newtonian fluids described by the Carreau constitutive model. The study covers both shear-thinning ($n < 1$) and shear-thickening ($n > 1$) regimes.

• Spectral Formulation: The authors derive spectral evolution equations for the variable viscosity system. They demonstrate that the variable viscosity term becomes nonlinear in spectral space, resulting in a convolution product formally analogous to the convective term.
• Coupling Analysis: Following the methodology of Couteau et al. [10] for convective transfer, the authors compute the viscous mode-to-mode coupling $\hat{V}(\kappa, \kappa')$ directly from the spectral equations. They decompose the total dissipation into a mean-viscosity contribution (classical local dissipation) and a fluctuating contribution arising from viscosity variations.
• Simulations: Simulations are performed on a triply periodic cubic domain ($512^3$ grid points) with random forcing. Cases include various power-law indices ($n$) and base viscosities to explore Reynolds number effects and scaling behaviors.

Key Contributions and Results

1. Nonlinear Viscous Coupling: The study establishes that the variable viscosity term generates a convolution product in spectral space, coupling all modes. Unlike the constant viscosity case, this term is not purely local. It entangles conservative (transfer) and non-conservative (dissipation) contributions, meaning the spectral quantity $\hat{V}(\kappa, \kappa')$ cannot be a priori separated into pure transfer and pure dissipation.

2. Viscous Energy Transfer: A central finding is that the nonlinear viscous term can act as a genuine inter-scale energy transfer mechanism.

   • Shear-Thickening Regime: In shear-thickening fluids, the viscous coupling $\hat{V}(\kappa, \kappa')$ exhibits a distinct "transfer-like dipole" structure near $\kappa' \approx \kappa$. This dipole satisfies the approximate antisymmetry $\hat{V}(\kappa, \kappa') \approx -\hat{V}(\kappa', \kappa)$, which is the defining signature of conservative energy transfer. This indicates that viscous effects actively redistribute energy between scales, a role traditionally attributed exclusively to convective nonlinearity.
   • Shear-Thinning Regime: In shear-thinning fluids, this dipole structure is absent. The viscous coupling remains dominated by non-conservative effects near the origin ($\kappa' \approx 0$), and the energy transfer is negligible compared to the convective term.

3. Spectral Budget and Dissipation Scaling:

   • Shear-Thickening: The emergence of viscous energy transfer in the dissipation range is linked to a fundamental change in spectral decay. While Newtonian and shear-thinning fluids exhibit the classical exponential cutoff in the dissipation range, shear-thickening fluids display a power-law decay $\hat{E}(\kappa) \sim \kappa^{-\alpha}$. The authors argue this algebraic scaling arises because the viscous flux sustains energy transfer to high wavenumbers, preventing the runaway dissipation that causes exponential cutoffs. The exponent $\alpha$ depends on both the power index $n$ and the Reynolds number, converging toward $\approx 7$ at higher Reynolds numbers for $n=3.0$.
   • Shear-Thinning: The exponential decay is preserved, consistent with the absence of significant viscous transfer.

4. Kolmogorov Scaling: Upon rescaling with the mean viscosity, the spectral fluxes for both shear-thinning and shear-thickening fluids collapse onto a single curve, confirming that the Kolmogorov cascade picture extends to these fluids, with the crossover to the viscous range occurring at $\kappa\eta_{eff} \approx 0.3$.

5. Mode-to-Shell Transfer: The viscous transfer function participates in the forward energy cascade alongside the convective term. In shear-thickening fluids, the viscous transfer grows from negligible at intermediate wavenumbers to dominant in the dissipation range, effectively taking over the role of the convective term in redistributing energy at small scales.

Significance and Claims
The paper claims that inter-scale energy transfer is not the exclusive province of the convective nonlinearity. It demonstrates that any nonlinear term in the momentum equation that produces a spectral convolution product can, in principle, carry conservative contributions. Whether such transfer is significant depends on the specific physics and must be assessed by direct computation of mode-to-mode coupling.

The authors highlight a broader implication for Large Eddy Simulation (LES). Since the Smagorinsky-Lilly closure models eddy viscosity as a power-law function of strain rate (effectively a shear-thickening fluid with $n=2$), the findings predict that in LES where eddy viscosity dominates molecular viscosity, the eddy-viscosity term should produce a viscous spectral flux in the dissipation range, leading to algebraic rather than exponential decay in the resolved energy spectrum. This provides a falsifiable prediction for spectral budget analysis in well-resolved LES.

Finally, the work clarifies that the entanglement of conservative and non-conservative effects in nonlinear spectral terms prevents separation based on physical equations alone; direct computation of the coupling is the only way to determine if a nonlinear term transfers energy or merely dissipates it.