🔢 mathematics

Delayed Dissipation for Two-Dimensional Vortex Sheets

This paper establishes significantly improved lower bounds on the energetic lifetime of two-dimensional Delort vortex sheets by proving that viscous energy dissipation is delayed for timescales up to polynomial or polylogarithmic orders in the inverse viscosity, depending on the regularity of the initial vorticity's non-positive part.

Original authors: Victor Armegioiu

Published 2026-08-04

📖 8 min read🧠 Deep dive

CC BY 4.0

Original authors: Victor Armegioiu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

=== DRAFT ===
Imagine a world where fluids, like water or air, are made of invisible, swirling tornadoes called "vortices." In the real world, these swirls eventually slow down and disappear because of "viscosity," which is just a fancy word for the fluid's internal stickiness or friction. But in the world of ideal physics, scientists often pretend this stickiness doesn't exist to make the math easier. The big question is: if we start with a fluid that has a sharp, jagged edge where the swirling speed jumps suddenly (like a vortex sheet), how long does it take for that stickiness to actually start stealing the fluid's energy?

For a long time, mathematicians knew that if you wait long enough, the energy will eventually vanish. But they were stuck on a tricky timeline. They knew that for a short while, the fluid behaves almost perfectly like the ideal, frictionless version. The mystery was: exactly when does the "friction" kick in to steal a noticeable amount of energy? Is it after a few seconds? A few years? Or does it take a time so long it's practically infinite? This paper dives into that specific timing, trying to figure out the "waiting time" before the fluid starts to lose its energy due to viscosity.

The Great Energy Heist: How Long Can a Swirl Hide?

Think of a fluid as a massive, energetic dance party. The dancers are the swirling vortices, and the music is their kinetic energy. In a perfect, frictionless world, they would dance forever without getting tired. But in our real world, there's a "sticky floor" (viscosity) that slowly steals the dancers' energy, turning their movement into heat.

The story starts with a specific type of dancer: a "vortex sheet." Imagine a line where the dancers on one side are spinning one way, and the dancers on the other side are spinning the other way, but right at the line, the speed jumps suddenly. It's a chaotic, jagged edge. The paper asks: If we have a crowd of these dancers, how long can they keep their energy before the sticky floor steals a significant chunk of it?

The Old Guess vs. The New Reality

A few years ago, other scientists (De Rosa and Marcotullio) made a guess. They thought that for these specific types of swirling lines, the energy theft would happen relatively quickly. They predicted that if you waited for a time related to the "stickiness" of the fluid (let's call it $\nu$ ), the energy loss rate would be noticeable after a time roughly proportional to $1/\sqrt{|\log \nu|}$ . In plain English, they thought the dancers would get tired and start losing energy fairly soon after the party started.

This paper says: "Not so fast."

The author, Victor Armengou, proves that the old guess was too optimistic. He shows that the dancers can hold onto their energy for much, much longer than previously thought. In fact, he proves that the energy loss rate is actually smaller than the old guess predicted. Instead of the loss being proportional to $1/\sqrt{|\log \nu|}$ , it is actually proportional to $1/|\log \nu|$ .

To put this in perspective, imagine the old guess said the dancers would get tired after running a mile. Armengou proves they can actually run a square root of a mile (which is a much longer distance) before they start to feel the burn. He explicitly rules out the idea that the energy loss happens as quickly as the previous researchers thought.

The Secret Weapon: The "One-Sided" Rule

How did he figure this out? He used a clever trick involving the "sign" of the swirls.

Imagine the dancers are divided into two groups: the "Positive Spinners" and the "Negative Spinners." In the specific type of fluid this paper studies, the chaotic, jagged edge (the vortex sheet) is made entirely of "Positive Spinners." They all spin in the same direction, even if they are mixed with some regular, messy background noise that spins both ways.

The previous scientists looked at the dancers in small, local neighborhoods. They asked, "What's the biggest group of dancers in this tiny circle?" and used that to estimate how fast energy would be lost. It was like judging the whole party's energy by looking at the loudest corner.

Armengou realized this was a mistake. Because all the "Positive Spinners" are on the same team, they interact with each other across the whole room. It's not just about the loudest corner; it's about the total energy of the entire team of Positive Spinners working together. He developed a new mathematical tool (an interpolation inequality) that looks at the total interaction of these like-signed dancers.

This new tool showed that the "Positive Spinners" are actually very good at holding onto their energy. They resist the sticky floor much better than the old method predicted. By keeping track of the total team effort rather than just the loudest local group, he found that the energy loss rate is slower by a factor of the square root of a logarithm.

The "Waiting Time" Surprise

The paper doesn't just say the energy loss is slower; it tells us how long we have to wait before any significant energy is lost.

However, there is a crucial condition: This long "waiting time" result applies specifically when the initial arrangement of the dancers is "relatively compact" in a mathematical sense. Think of this as the dancers starting in a well-organized, predictable formation that doesn't have wild, chaotic high-frequency jitters.

If this condition is met, the paper proves that you have to wait at least until the time is a polynomial function of the inverse of the stickiness ( $1/\nu$ ).

Think of it this way:

The old guess said you'd have to wait for a time that grows like a stretched exponential (very fast growth, but still finite).
The new proof says that for these well-organized starting points, you have to wait for a time that grows like a power of the stickiness (like $1/\nu$ raised to some power).

This is a huge difference. It means the "inviscid" (frictionless) model is accurate for a much longer time than anyone thought. The fluid can pretend it has no friction for a very long time before the real physics of stickiness kicks in to steal the energy.

But what if the starting point isn't well-organized? If the initial velocities are not relatively compact (meaning they have wild, high-frequency variations that change as the fluid gets less sticky), the paper shows that the energy loss can happen faster. In fact, for a single, fixed radial setup, the energy loss is even faster than the polynomial delay, vanishing strictly faster than $1/|\log \nu|$ . So, the "super-long" waiting time is a special feature of well-behaved, compact starting conditions.

The "Radial" Exception

The paper also checks what happens if the dancers are arranged in perfect circles (radial symmetry). Here, the story is a bit different.

If you have a family of different starting setups that change as the fluid gets less sticky (and these setups are not relatively compact), you can still get the slower energy loss rate ( $1/|\log \nu|$ ).
But if you pick one single, fixed starting setup (a fixed radial dance routine), the energy loss is even faster! It vanishes strictly faster than $1/|\log \nu|$ .

This means that for a single, unchanging circular swirl, the "sticky floor" starts stealing energy almost immediately on the diffusive scale. The "delayed dissipation" only happens when you have a specific, carefully constructed family of changing starting points.

The Bottom Line

This paper is a mathematical proof, not a simulation or a guess. It uses rigorous logic to show that:

The energy loss is slower than previously believed for these specific types of swirling fluids.
The "waiting time" before energy is lost is much longer (polynomial) than the old "stretched exponential" guess, but only if the initial fluid motion is relatively compact.
The reason is that the "one-sided" nature of the swirls (all spinning the same way) creates a collective resistance to friction that was previously overlooked.

So, the next time you see a fluid swirl, remember: if it's a jagged edge of like-signed dancers with a well-organized start, they are tougher than they look. They can hold onto their energy for a surprisingly long time, defying the sticky floor for much longer than the old rules of the game suggested. The "energy heist" is happening, but the thieves are moving in slow motion.

Technical Summary: Delayed Dissipation for Two-Dimensional Vortex Sheets

Problem Statement
The paper addresses the quantitative problem of viscous energy loss in two-dimensional Navier–Stokes flows with initial data in the Delort class. Specifically, it investigates the rate at which kinetic energy is dissipated by viscosity for vortex sheets where the vorticity contains a singular, non-negative component (a "distinguished-sign" measure) and a regular, possibly sign-changing component. While previous qualitative results established that such flows do not exhibit anomalous dissipation on fixed time intervals as viscosity $\nu \to 0$ , the precise dependence of the dissipation rate on $\nu$ and the time scale $T$ remained open. The paper focuses on disproving a specific conjecture by De Rosa and Marcotullio regarding the sharpness of the dissipation bound and determining the minimal time required for a fixed amount of energy to be dissipated.

Methodology
The core of the analysis relies on a sharp interpolation inequality for non-negative densities in $H^{-1}$ and $L^2$ spaces, combined with the enstrophy identity.

Sharp Logarithmic Interpolation: The author derives a new $L^2$ – $H^1$ – $H^{-1}$ interpolation inequality for non-negative measures. Unlike previous arguments that relied on "worst-ball" estimates (which bound the mass in the largest small ball), this method retains the total interaction energy of the positive vorticity. By exploiting the positivity of the Bessel kernel (the Green's function for $1-\Delta$ ), the author shows that the total logarithmic interaction of close, like-signed pairs provides a coercive lower bound on the enstrophy. This yields a logarithmic gain in the interpolation estimate, improving the denominator from a fourth root of a logarithm to a square root.
Propagation of Decomposition: The paper demonstrates that the Delort decomposition ( $\omega = \mu + f$ , with $\mu \geq 0$ ) is preserved under the Navier–Stokes flow. The positive measure component $\mu$ and the regular component $f$ evolve with uniform bounds independent of viscosity, provided the initial data satisfies specific integrability conditions (e.g., $f \in L^p$ or $f \in L(\log L)^\alpha$ ).
Enstrophy Identity and Integration: By combining the sharp interpolation inequality with the enstrophy identity ( $\frac{d}{dt}\|\omega\|_2^2 = -2\nu \|\nabla \omega\|_2^2$ ), the author derives differential inequalities for the $L^2$ norm of the vorticity. Integrating these inequalities yields explicit bounds on the cumulative energy dissipation $D_\nu(\delta, T)$ .
Exact Radial Solutions: To establish the sharpness of the derived upper bounds, the paper constructs exact radial solutions on $\mathbb{R}^2$ . These solutions utilize heat kernel evolutions, where the transport term vanishes due to radial symmetry, allowing for precise calculation of dissipation rates for various regularity classes of the initial data.

Key Contributions and Results

Disproof of Conjecture 1.6: The paper disproves the conjecture by De Rosa and Marcotullio [6] that there exists a Delort datum for which the dissipation $D_\nu(\delta, 1)$ scales as $O(|\log \nu|^{-1/2})$ . The author proves that for any datum in the conjectured class (finite energy, $L^p$ regular part, non-negative singular part), the dissipation is strictly bounded by $O(|\log \nu|^{-1})$ .
Improved Dissipation Rate: The main theorem establishes that for fixed $0 < \delta < T$ , the energy loss satisfies:
$D_\nu(\delta, T) \lesssim \frac{1}{|\log \nu|}$
This improves the previous $O(|\log \nu|^{-1/2})$ bound.
Delayed Dissipation and Waiting Times: Under the assumption that initial velocities are relatively compact in $L^2$ , the paper shows that energy loss is delayed significantly. Specifically, if the observation time $T_\nu$ grows such that $\log T_\nu = o(|\log \nu|)$ , the dissipation vanishes as $\nu \to 0$ . This implies that any prescribed energy loss $\eta$ must wait until a time scale of at least $\nu^{-a}$ for some $a > 0$ , providing a polynomial lower bound on the energetic lifetime of the inviscid model.
Orlicz Space Generalization: The results are extended to cases where the regular part of the vorticity belongs to Orlicz spaces $L(\log L)^\alpha$ . The dissipation rate is shown to be $O(|\log \nu|^{-q_\alpha})$ where $q_\alpha = \min\{2\alpha, 1\}$ . The threshold $\alpha = 1/2$ corresponds to the embedding $L(\log L)^{1/2} \subset H^{-1}$ , beyond which the rate saturates at $O(|\log \nu|^{-1})$ .
Sharpness via Radial Solutions:
- On $\mathbb{R}^2$ , exact radial solutions attain the positive-time exponents $q_\alpha$ for $0 < \alpha \leq 1/2$ .
- A bounded-energy family without relative compactness attains the endpoint rate $O(|\log \nu|^{-1})$ .
- Conversely, for any fixed radial datum in the finite-energy endpoint class, the dissipation is strictly $o(|\log \nu|^{-1})$ , and the energy loss occurs only on the diffusive scale $1/\nu$ .

Significance and Claims
The paper claims that the "distinguished-sign" hypothesis (one-signed singular vorticity) fundamentally limits the speed at which vorticity can concentrate enough enstrophy to allow viscosity to remove kinetic energy. The improvement from a fourth-root to a square-root logarithmic gain in the interpolation inequality is identified as the mechanism that tightens the dissipation bound and extends the time scale of validity for the inviscid limit.

The author emphasizes that their results clarify the time-scale question left open by previous qualitative theories: while anomalous dissipation does not occur on fixed intervals, the onset of significant energy loss is delayed by a power of the Reynolds number. The paper explicitly states that while the positive-time exponents are shown to be sharp via radial examples, the waiting-time lower bounds are not proven to be attained by general fixed data, and the question for non-radial fixed data remains open. The work serves as a quantitative refinement of the no-anomalous-dissipation theory for Delort vortex sheets.