Boundary-layer separation beyond the square-root law

arXiv:2610.01689 · math.AP · Submitted 2026-10-01 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: We're looking at the paper: "Boundary-layer separation beyond the square-root law".

Emmy: Here's the gist of the paper. The paper constructs solutions to the two-dimensional stationary Prandtl system under a constant adverse pressure gradient, revealing a family of distinct separation laws indexed by an integer l at least 1.

Ada: For every l at least 1, smooth initial data is constructed such that separation occurs at a finite location x=x*, with the boundary condition d y u(x, 0) = C l(x* - x) l/2 + o(1) as x x*.

Emmy: The local asymptotic structure near separation is determined using matched inner and outer expansions based on linearization around the ground state in von Mises variables.

Ada: Theorem 1.1 establishes the dynamic stability of the square-root separation regime (l=1).

Paper summary: Emmy: So, wrapping up this discussion on "Boundary-layer separation beyond the square-root law," the authors have shown that for every integer l greater than or equal to one, you can construct smooth initial data for the two-dimensional stationary Prandtl system under a constant adverse pressure gradient that exhibits a family of distinct separation laws.

Ada: They've established that they can find separation points x star l for every positive x zero, and for l greater than or equal to one, the wall shear stress near that point follows the relationship involving C l and t raised to the power of one half plus a small error term that decays with t.

Kolya: And they also proved that the square-root separation regime, which corresponds to l equal to one, is dynamically stable under small admissible perturbations of the initial data in a suitable topology.

Emmy: This work suggests that separating boundary layers aren't just one type of problem, but an entire family dictated by this integer index l. The authors achieved this by using sophisticated techniques like matching formal defect hierarchies with spectral expansions and modulation equations to determine the local asymptotic structure near separation for every l.

Ada: It’s about showing that you can systematically construct solutions for these different separation regimes, and it implies that the physical situation isn't as simple as just one curve. The paper confirms that this family of higher-order laws can be constructed through finite-dimensional tuning.

Conclusion: Ada: So, we've been looking at this paper about boundary layer separation and how it goes beyond that old square root law for different orders of separation, l.

Emmy: Yeah, the title itself suggests they're going beyond just one simple case. They’re looking at a whole family of these different separation laws indexed by this integer l.

Kolya: What I want to focus on is that they constructed solutions for every single integer l greater than or equal to one. That's a lot of distinct physical scenarios they managed to solve.

Ada: Exactly, and the authors showed that for each l, there's a specific way the wall shear stress behaves right at that separation point, following this formula involving C l and the distance to it raised to the power of one half.

Emmy: And they didn't just stop at finding these solutions; they proved dynamic stability for every case, even for that basic square root case where l is one. That's a big deal when you're dealing with fluid dynamics.

Kolya: For the higher orders, l greater than or equal to two, they showed those can be constructed through this countable family of finite-dimensional tuning, which means it’s not just some weird fluke; there’s a systematic way to find them.

Ada: It means that instead of one universal separation mechanism, you have this whole spectrum of different behaviors depending on the pressure gradient and the geometry.

Emmy: And this work suggests that understanding these different l values gives us a much richer picture of how these boundary layers actually break down.

Ky Giao Duong, Tej-Eddine Ghoul, Sameer Iyer, Yasunori Maekawa

Department of Mathematics, NYU Abu Dhabi · Department of Mathematics, NYU Abu Dhabi · Department of Mathematics, University of California Davis Department of Mathematics Graduate School of Science Kyoto University

math.AP

Submitted: 2026-10-01

Updated: 2026-11-06

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 88/100

The gist: The paper constructs solutions to the two-dimensional stationary Prandtl system under a constant adverse pressure gradient…

Key concepts

Prandtl system
This refers to a mathematical model describing how fluid velocity changes near a solid surface, specifically in boundary layers. The paper studies this system when there is a constant adverse pressure gradient, which typically causes the flow to separate from the wall.
$l$-indexed separation laws
The paper finds that separation does not happen in just one way; instead, it reveals a family of distinct separation laws indexed by an integer $l imes 1$. This means there are different mathematical forms describing how the flow separates, with each form corresponding to a specific integer value of $l$.
Von Mises variables
These are special coordinates used in the analysis, specifically when linearizing around the ground state of the system. They help simplify complex fluid dynamics problems by transforming them into a simpler form where asymptotic structures can be determined using matched expansions.
Local asymptotic structure
This describes the mathematical shape of the flow very close to where separation occurs. The paper uses inner and outer expansions, based on these variables, to show that near the separation point $x^*$, the solution looks like a specific function involving powers of $(x^* - x)$.

Terminology

Summary

The paper constructs solutions to the two-dimensional stationary Prandtl system under a constant adverse pressure gradient, revealing a family of distinct separation laws indexed by an integer l at least 1. p. 0 For every l at least 1, smooth initial data is constructed such that separation occurs at a finite location x=x*, with the boundary condition d y u(x, 0) = C l(x* - x) l/2 + o(1) as x x*. p. 0 The local asymptotic structure near separation is determined using matched inner and outer expansions based on linearization around the ground state in von Mises variables. p. 0

Theorem 1.1 establishes the dynamic stability of the square-root separation regime (l=1). p. 3 Specifically, for every x 0 > 0, a solution u bin,1 is constructed that separates at x b* 1 in (0, x 0/2). p. 3 The local expansion near this point is given by:

u 1(x, y) = C b2 1 t Q sep y C b1 t 1/2 + u e1(x, y)

where t = x b* 1 - x, and the remainder term is bounded by u e1(x, y) t 1/2 + eta y + y cubed for 0 at most x at most x b* 1 and 0 at most y at most y 0.

Theorem 1.2 extends this result to a countable family of higher-order separation laws for l at least 2. p. 4 For every integer l at least 2 and every x 0 > 0, a solution u is constructed that separates at x* in (0, x 0). p. 4 The local asymptotic structure near this point is described by:

u(x, y) = C 2 t l Q sep y C b t l/2 + u e(x, y)

where t = x* - x, and the remainder term is bounded by u e(x, y) t l/2 + (2l-1) eta y + y 2l+1 for 0 at most x at most x* and 0 at most y at most y 0.

The construction of these solutions involves matching the formal defect hierarchy with the spectral expansion around the ground state in von Mises variables, leading to an adapted decomposition. p. 5 The separation point x* is controlled by preparation parameters, allowing it to be prescribed near the inlet: for each fixed l at least 1 and x 0 > 0, there exists a function b x*(x 0) such that every separation point x* in (0, b x*(x 0)) is attained. p. 4 For l at least 2, this is achieved by setting the separation point as x* = e-s 0 in a specific parameterization. p. 2

The analysis employs modulation equations to determine the scale and profile of the solution, which are related to modulation–energy constructions for Type II dynamics and modulation–comparison methods used for Prandtl separation. p. 5 The construction yields a global renormalized solution of (Pstab) in (3. p. 43). The final result confirms that for every l at least 1, a separation point x* exists such that the wall shear stress satisfies d y u(x, 0) = C(x* - x) 1/2(1 + O((x* - x) eta)) as x to (x*)-, ensuring separation at the finite point x*. p. 2 For l at least 2, Theorem 1. p. 02 demonstrates that a countable family of higher-order separation laws can be constructed through finite-dimensional tuning. p. 4

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Enhanced prediction of boundary layer separation dynamics: The system can now predict distinct separation laws for every integer l greater than or equal to 1, including a family of higher-order laws for l greater than or equal to 2, based on the initial data structure.

  2. Rigorous analysis of stability in the square-root regime: The improved system can determine if a specific separation law (like the one corresponding to l=1) is dynamically stable under small perturbations of the initial data within a suitable topology.

  3. Construction of sharp local asymptotic structures: The AI can determine the local asymptotic structure of solutions near separation by constructing matched inner and outer expansions based on linearization around the ground state in von Mises variables.

  4. Quantitative characterization of separation profiles: It can provide explicit formulas for the solution near separation, such as u1(x, y) = Cb 2/1 t Qsep y Cb1t 1/2, allowing for quantitative prediction of how the separation point moves and how the wall shear behaves.

  5. Robust initial data generation: The system can construct smooth initial data for which separation occurs at a finite location x star with a specified initial slope behavior, ensuring physical admissibility and adherence to Oleinik's local theory.

  6. Stability analysis of higher-order regimes: For l greater than or equal to 2, the AI can analyze the unstable regimes by reducing the problem to a finite-dimensional system and identifying the specific directions where instabilities arise in the expansion about the ground state.

  7. Adaptive parameter selection for flow control: The system can select separation point parameters (like b and beta) through geometric orthogonality conditions, enabling a refined analysis of how physical constraints govern flow behavior.

Sources

Related papers