Uniqueness problem for Prandtl spirals

Tomasz Cieślak, Piotr Kokocki, Przemysław Kosewski · 2025

Read Uniqueness problem for Prandtl spirals by Tomasz Cieślak, Piotr Kokocki, Przemysław Kosewski (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-03554.txt.

Introduction

A Prandtl spiral, a swirling flow whose vorticity winds outward in a logarithmic coil, looks simple, but is the velocity field behind it the only possible one? Tomasz Cieślak, Piotr Kokocki, and Przemysław Kosewski settle the question in this 2025 paper. Under a matching condition at the spiral and a decay condition at its center, the field is uniquely fixed.

They extend the verdict to vorticity built from several concentric logarithmic spirals, and recover the classic velocity formula along the way. The proof's engine is a conformal map that unwraps the spiral's exterior into a strip, turning the problem into one about holomorphic functions with unusual boundary rules.

Catalog record

CREATORTomasz Cieślak, Piotr Kokocki, Przemysław Kosewski
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE72 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-03554

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Uniqueness problem for Prandtl spirals Tomasz Cieślak, Piotr Kokocki, Przemysław Kosewski Abstract . This paper is devoted to the study of the following problem: Is the divergence-free velocity field introduced in [ 3 ] the unique such field associated with vorticity given by the Prandtl spiral? Without additional assumptions, the answer is negative. However, we show that if the class of admissible velocities is restricted to those satisfying the velocity matching condition and an appropriate decay condition at the origin of the spiral, then the velocity field is uniquely determined. We subsequently extend the result to the case of fields with vorticity composed of unions of concentric logarithmic spirals. As a by-product, we derive an alternative way of deriving formula for the velocity corresponding to the Prandtl spirals. The proof relies on an approach that is of independent interest. We construct an explicit conformal map from the exterior of a logarithmic spiral onto a strip. This transformation reduces the problem to establishing the uniqueness of a holomorphic function defined on the strip, under non-standard boundary and decay conditions. 1991 Mathematics Subject Classification 76M40, 76B47 1. Introduction The velocity field corresponding to the vorticity induced by a single Prandtl spiral, as well as by a set of N\geq 2 uniformly distributed such spirals, was explicitly constructed in [ 3 ] . Recall that the Prandtl spirals are well-known objects in hydromechanics, postulated by Helmholtz himself in his foundational paper on discontinuous flows [ 9 ] . They were considered due to the observability of spiraling flows in nature and the conjecture – tracing back to Helmholtz – that such structures are inherently unstable and may give rise to further instabilities, such as the well-known von Kármán vortices. Many authors have investigated flows exhibiting spiral patterns – see, for instance, Hamel [ 8 ] . A well-known example of logarithmic spiral vortex sheets was introduced by Prandtl [ 11 ] , who associated them with the flow detachment from the wingtip of a plane during takeoff. Alexander later extended Prandtl’s construction to several uniformly distributed spirals; see [ 2 ] . For more recent observations of spiral-patterned flows, see [ 7 ] ; for applications in aviation, see [ 13 ] . While Prandtl spirals have appeared in the engineering literature for over a century, their rigorous and exact mathematical status was not well understood until very recently. First, Elling and Gnann [ 5 ] observed that a family of N\geq 3 uniformly distributed logarithmic spirals satisfies the Birkhoff-Rott equation. Subsequently, the first two authors of the present article, together with Ożański, showed that not only a single Prandtl spiral, but also any set of N\geq 2 uniformly distributed logarithmic spirals, serves a support of the vorticity field that gives rise to velocity field solving the weak 2D Euler equation, see [ 3 ] . In a moment, we shall review the main points of this construction. Let us only recall that bifurcation branches of non-symmetric logarithmic spirals solving the 2D Euler equation was constructed in [ 4 ] . An alternative construction of the Prandtl spiral as a weak solution to the 2D Euler equation was provided by Jeong and Said in [ 10 ] , where the results of [ 4 ] were also independently rederived using a different approach. Let us recall that the Prandtl spiral is a time-dependent flow with a vorticity concentrated on a family of curves \{\Sigma(t)\}_{t>0} , parameterized by (1.1) Z(\theta,t)=t^{\mu}e^{a\theta}e^{i\theta},\quad t>0, where \theta\in\mathbb{R} is an angular parameter and a>0 , \mu\in\mathbb{R} are constants. Moreover, the total circulation along the spiral is given by \Gamma(\theta,t)=gt^{2\mu-1}e^{2a\theta},\quad t>0, where g\in\mathbb{R}\setminus\{0\} is a constant. In [ 3 , eq. (1.12)] , the following formula for the velocity field associated with the Prandtl spiral was derived: (1.2) v(z,t)=t^{\mu}w(z/t^{\mu}),\quad z\in\mathcal{R}(t),\ t>0, where \mathcal{R}(t):=\mathbb{C}\setminus(\Sigma(t)\cup\{0\}) is the time-dependent complement of the spiral and the profile function is expressed in the polar coordinates as (1.3) w(z)=e^{i\theta}\frac{2ag}{r(a-i)}\left(r^{\frac{2a}{a+i}}e^{-A\theta}\frac{e^{2\pi J(r,\theta)A}}{1-e^{2\pi A}}\right)^{*} for z=re^{i\theta}\in\mathcal{R}:=\mathcal{R}(1) . In the above formula A:=-2ai/(a+i) and J(r,\theta) is a winding number of the spiral \Sigma:=\Sigma(1) , given by (1.4) J(r,\theta)=\min\{j\in\mathbb{Z}\ |\ a(2\pi j-\theta)+\ln r>0\}. We know that the profile function ( 1.3 ) can be recovered as the derivative of a suitable complex potential. In fact, one can show that w^{*}=\Phi^{\prime} , where (1.5) \displaystyle\Phi(z)=\frac{g}{1-e^{2\pi A}}e^{iA(\ln r+i(\theta-2\pi J(r,\theta)))}, for z=re^{i\theta}\in\mathcal{R} . In [ 3 , Theorem 1.8] , it was shown that v is a weakly divergence-free vector field satisfying (1.6) \mathrm{curl}\,v(t)=\gamma(t)\,\delta_{\Sigma(t)},\quad t>0 in the distributional sense, where the measure density is given by (1.7) \gamma(t,Z(t,\theta)):=\frac{\partial_{\theta}\Gamma(t,\theta)}{|\partial_{\theta}Z(t,\theta)|}=\frac{2agt^{\mu-1}}{\sqrt{1+a^{2}}}e^{a\theta},\quad\theta\in\mathbb{R}. Moreover, the velocity profile satisfies the decay condition at the origin (1.8) |w(z)|\lesssim|z|,\quad z\to 0,\ z\in\mathcal{R}. A natural question then is whether the time dependent velocity field defined by ( 1.2 ), ( 1.3 ) is the only one on \mathbb{R}^{2} whose vorticity matches ( 1.6 ) and ( 1.7 ), and whose decay at the origin satisfies ( 1.8 ). This question is both significant and nontrivial for the following reasons. Irregularity of the Prandtl-spiral vorticity. The vorticity associated with the Prandtl spiral is so singular that the Biot–Savart integral formula fails to converge, preventing direct use of the usual integral representation. Non-uniqueness up to holomorphic additions. Away from the spiral itself, the flow is irrotational, so adding any holomorphic function to w leaves the vorticity unchanged. Therefore, additional constraints on the class of velocity fields must be imposed to rule out such “hidden” degrees of freedom. Linear growth at infinity. Our method does not impose any growth‐type restrictions on w . Since the velocity field grows linearly at infinity, no Liouville type argument (which would require boundedness) can be applied. Following [ 3 ] , we focus on the self-similar logarithmic spiral vortex sheets, for which the profile map w satisfies the continuity condition for normal components (1.9) (w^{R}(z)-w^{L}(z))\cdot\vec{n}(z)=0,\quad z\in\Sigma, and the tangential jump condition (1.10) (w^{R}(z)-w^{L}(z))\cdot\vec{\tau}(z)=\gamma(z),\quad z\in\Sigma, where w^{R}(z) and w^{L}(z) are the limit velocities w(z^{\prime}) as z^{\prime}\to z from the right and left sides of the spiral \Sigma with respect to its natural orientation. Moreover, \vec{\tau}(z) and \vec{n}(z):=-i\vec{\tau}(z) are the unit tangent and normal vectors to the spiral, respectively, and \gamma:=\gamma(1) is a density function. According to the formula ( 1.1 ), the spiral \Sigma is parametrized by the mapping Z(\theta):=Z(\theta,1)=e^{a\theta}e^{i\theta},\quad\theta\in\mathbb{R} and the corresponding unit tangent and normal vectors are given by (1.11) \displaystyle\vec{\tau}(Z(\theta)) \displaystyle=\frac{1}{\sqrt{a^{2}+1}}[a\cos\theta-\sin\theta,\cos\theta+a\sin\theta], (1.12) \displaystyle\vec{n}(Z(\theta)) \displaystyle=\frac{1}{\sqrt{a^{2}+1}}[\cos\theta+a\sin\theta,\sin\theta-a\cos\theta]. Furthermore, the density function satisfies (1.13) \gamma(Z(\theta))=\frac{\partial_{\theta}\Gamma(\theta)}{|\partial_{\theta}Z(\theta)|}=\frac{2ag}{\sqrt{1+a^{2}}}e^{a\theta},\quad\theta\in\mathbb{R}, where \Gamma(\theta):=\Gamma(\theta,1) . Let us emphasize that ( 1.9 ) and ( 1.10 ) represent the standard jump conditions for vortex sheets. They are, respectively, equivalent to the condition that the velocity profile w is divergence-free and that its vorticity satisfies ( 1.6 ) in the sense of distributions. Moreover, as we point out in [ 3 , Theorem 1.2] , if the self-similar vector field v is a weak solution of the 2D Euler equation, then the profile function satisfies the velocity matching condition (1.14) \vec{n}(z)\cdot\left(w(z)-\mu z\right)=0,\quad z\in\Sigma. Observe that condition ( 1.14 ) is well-defined since, by ( 1.9 ), we can set \vec{n}(z)\cdot w(z):=\vec{n}(z)\cdot w^{L}(z)=\vec{n}(z)\cdot w^{R}(z) for z\in\Sigma . We are now ready to formulate our main theorem and discuss the strategy of its proof. Theorem 1.1 . Let w be a two-dimensional weakly

Read the complete text