Asymmetric Self-similar Spiral Solutions of 2-D Incompressible Euler Equations
Hyungjun Choi · 2025
Read Asymmetric Self-similar Spiral Solutions of 2-D Incompressible Euler Equations by Hyungjun Choi (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-2507-05059.txt.
Introduction
Spiraling flows that look the same at every scale are prized test cases for fluid mathematics, but known examples all repeat themselves around the circle. Hyungjun Choi removes that crutch. Working with the two-dimensional incompressible Euler equations, which govern ideal fluids, Choi constructs self-similar algebraic spiral solutions with no rotational symmetry at all. Earlier work by Elling and by Shao, Wei, and Zhang had built such spirals only under an assumed m-fold symmetry.
The result extends the family of exact spiral flows beyond the symmetric cases. Such solutions serve as benchmarks for theories of how spiral patterns in fluids form, sharpen, and persist.
Catalog record
| CREATOR | Hyungjun Choi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 101 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-05059
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Asymmetric Self-similar Spiral Solutions of 2-D Incompressible Euler Equations Hyungjun Choi Abstract We construct nonradial, self-similar solutions to the two-dimensional incompressible Euler equations without assuming rotational symmetry. These solutions extend the study of self-similar algebraic spiral flows, initiated by Elling and further developed by Shao–Wei–Zhang [ 41 ] , where m m -fold symmetry with m ≥ 2 m\geq 2 was assumed. Moreover, they bear resemblance to the numerical simulations of Bressan-Shen [ 10 ] , in connection with the ongoing investigation into non-uniqueness of solutions. 1. Introduction We study solutions of the 2-D incompressible Euler equations \left|\begin{array}[]{rcl}\partial_{t}u+u\cdot\nabla u+\nabla p&=&0,\\ u(t=0)&=&u_{0}.\end{array}\right. (1.1) Considering the vorticity \omega=-\partial_{2}u_{1}+\partial_{1}u_{2} , we obtain the vorticity equation: \left|\begin{array}[]{rcl}\partial_{t}\omega+u\cdot\nabla\omega&=&0,\\ u=\nabla^{\perp}\psi,\penalty\ \Delta\psi&=&\omega,\\ \omega(t=0)&=&\omega_{0},\end{array}\right. (1.2) where \psi is the stream function and \nabla^{\perp}=(-\partial_{2},\partial_{1})^{t} . Since \omega is transported by a divergence-free vector field u , the distribution of \omega is conserved in time. In particular, \lVert\omega(t,\cdot)\rVert_{L^{\infty}}=\lVert\omega_{0}\rVert_{L^{\infty}} . This, together with the Beale-Kato-Majda continuation criterion, guarantees the global existence of smooth solutions. On the other hand, the borderline spaces \omega\in L^{\infty} or u\in C^{1} are ill-posed; see [ 8 , 22 ] . We now turn to the study of non-smooth solutions. A distribution u(t,x)\in L^{2}_{\text{loc}}([0,\infty)\times\mathbb{R}^{2};\mathbb{R}^{2}) is called a weak solution to the Euler equation ( 1.1 ) with initial data u_{0} if \int_{\mathbb{R}^{2}}u_{0}\phi(t=0)\,dx+\int_{0}^{\infty}\int_{\mathbb{R}^{2}}(u\otimes u):\nabla\phi\,dx\,dt=0, for all divergence-free test functions \phi(t,x)\in C^{\infty}_{c}([0,\infty)\times\mathbb{R}^{2};\mathbb{R}^{2}) . A distributional weak solution \omega to ( 1.2 ) is defined analogously. Yudovich [ 50 ] proved that there is a unique weak solution \omega\in L^{\infty}(\mathbb{R};L^{1}\cap L^{\infty}(\mathbb{R}^{2})) , provided \omega_{0}\in L^{1}\cap L^{\infty} and u_{0}\in L^{2} . The existence of weak solutions has been extended to broader classes of initial data. DiPerna and Majda [ 19 ] showed the existence of the weak solution of 2-D incompressible Euler equations with initial data \omega_{0}\in L^{1}\cap L^{p} , u_{0}\in L^{2}_{\text{loc}} that conserves \lVert\omega\rVert_{L^{1}\cap L^{p}} . Later, Delort [ 17 ] extended the existence result to bounded, distinguished sign measure \omega_{0}\in\mathcal{M}\cap H^{-1}_{\text{loc}} , which includes the vortex sheet case. For the uniqueness, L^{\infty} -control of the vorticity can be relaxed to the so-called Yudovich space [ 51 ] . See [ 46 , 7 , 6 ] for further developments on the well-posedness of weak solutions with unbounded vorticity. The integrability condition of the vorticity can also be altered. Benedeto-Marchioro-Pulvirenti [ 5 ] showed the well-posedness of the class \omega_{0}\in L^{p}\cap L^{\infty} , |u_{0}|\lesssim 1+|x|^{\alpha} , for p<2/\alpha , \alpha<1 . While Serfaty [ 40 ] established the well-posedness of the class \omega_{0},u_{0}\in L^{\infty} , without any integrability condition. More recently, Elgindi-Jeong [ 21 ] showed that \omega_{0}\in L^{\infty} with m -fold rotational symmetry for m\geq 3 is sufficient for the well-posedness. In this work, we are interested in self-similar solutions of the 2-D incompressible Euler equation, which take the form \omega(t,x)=t^{-1}\Omega(t^{-\mu}x),\quad u(t,x)=t^{-1+\mu}U(t^{-\mu}x). (1.3) The vorticity profile \Omega is singular at the origin, is locally integrable, and could be sign-changing. The initial data is (-1/\mu) -homogeneous, placing it outside the scope of the existence and uniqueness results discussed above. In the remainder of the introduction, we review previous work in this direction and present our main result. These self-similar spiral solutions are potentially related to non-uniqueness phenomena in the 2-D incompressible Euler equations in two distinct contexts, which we elaborate on in section 1.2 . 1.1. Self-similar spiral solutions and main result Algebraic spirals naturally arise from self-similar solutions of the 2-D incompressible Euler equations. If a self-similar profile ( 1.3 ) solves the vorticity equation ( 1.2 ), then \Omega satisfies \left|\begin{array}[]{rcl}(U-\mu x)\cdot\nabla\Omega&=&\Omega,\\ U&=&\nabla^{\perp}\Delta^{-1}\Omega.\end{array}\right. (1.4) Since the radial vortex \omega(x)=\omega(|x|) is a steady state solution to ( 1.2 ), the function \overline{\Omega}(x)=C|x|^{-\frac{1}{\mu}}, is a solution to ( 1.4 ). The characteristic curve of the equation is \theta=\theta_{0}+C(2-\frac{1}{\mu})^{-1}r^{-\frac{1}{\mu}} , where (r,\theta) is the polar coordinates. This illustrates how algebraic spirals emerge as perturbations of \overline{\Omega} . In a series of works [ 23 , 24 ] , Elling established the following: Theorem 1.1 ( [ 24 ] , Theorem 1) . Given \mu>\frac{2}{3} and M>0 , there is N_{0}\in\mathbb{Z}_{+} such that a weak solution u of ( 1.1 ) exists with initial data \omega_{0}=r^{-\frac{1}{\mu}}\mathring{\omega}(\theta) , whenever \mathring{\omega} is N -fold rotationaly symmetric for N\geq N_{0} and \sum_{n\neq 0}|\widehat{\mathring{\omega}}_{n}|\leq M|\widehat{\mathring{\omega}}_{0}|. This result was generalized by Shao-Wei-Zhang [ 41 ] , who relaxed the symmetry assumption and extended the range of the parameter \mu . Theorem 1.2 ( [ 41 ] , Theorem 1.2) . Given \mu>\frac{1}{2} and N\in\mathbb{Z}_{>1} , there is an \epsilon>0 , independent of m , such that a weak solution u of ( 1.1 ) exists with initial data \omega_{0}=r^{-\frac{1}{\mu}}\mathring{\omega}(\theta) , whenever \mathring{\omega} is N -fold rotationaly symmetric and \lVert\mathring{\omega}-\widehat{\mathring{\omega}}_{0}\rVert_{L^{1}}\leq\epsilon N^{\frac{1}{2}}|\mathring{\omega}_{0}|. Our main result generalizes the above theorems by completely removing the rotational symmetry assumption. Main Theorem . Given \mu>1 , there exists \epsilon>0 such that for any \mathring{\omega}\in L^{1}(\mathbb{T}=\mathbb{R}/2\pi\mathbb{Z}) satisfying \sum_{n\in\mathbb{Z}\setminus\{0\}}|n|^{-\frac{1}{2}}|\widehat{\mathring{\omega}}_{n}|\leq\epsilon|\widehat{\mathring{\omega}}_{0}|, (1.5) there is a weak solution u\in C([0,\infty);L^{2}_{\text{loc}}) to ( 1.1 ) and \omega\in C([0,\infty);L^{1}_{\text{loc}}) to ( 1.2 ) where • (initial data) \omega_{0}(x)=r^{-\frac{1}{\mu}}\mathring{\omega}(\theta) , u_{0}(x)=\nabla^{\perp}\Delta^{-1}\omega_{0} ; • (self-similar) \omega(t,x)=t^{-1}\Omega(t^{-\mu}x) for \Omega\in L^{1}_{\text{loc}}(\mathbb{R}^{2}) , and u(t,x)=t^{-1+\mu}U(t^{-\mu}x) for a continuous div-free vector field U . Remark 1 . The initial stream function \psi_{0}=\Delta^{-1}\omega_{0} could be found via Fourier series. Since 2-\frac{1}{\mu} is never an integer, the equation \Big(2-\frac{1}{\mu}\Big)^{2}\mathring{\psi}+\partial_{\theta}^{2}\mathring{\psi}=\mathring{\omega}, has a unique solution \mathring{\psi} . Setting \psi_{0}=r^{2-\frac{1}{\mu}}\mathring{\psi}(\theta) yields \Delta\psi_{0}=\omega_{0} . Remark 2 . Our choice of space for \mathring{\omega} is motivated by the Wiener algebra setting used in [ 24 ] . Similar results with assumption ( 1.5 ), but still with m -fold rotational symmetry with large m , were obtained in [ 29 ] . The novelty of our result lies in eliminating this symmetry via improved linear analysis. Remark 3 . When N -fold rotational symmetric data is considered, our result recovers that of [ 41 ] (see Theorem 1.2 above), in a different smallness condition: \sum_{k\neq 0}|k|^{-\frac{1}{2}}|\widehat{\mathring{\omega}}_{kN}|\leq\epsilon N^{\frac{1}{2}}|\widehat{\mathring{\omega}}_{0}|. Since the smallness parameter \epsilon>0 is independent of N , taking N large can serve as a substitute for the smallness assumption, consistent with the approach in [ 24 ] . Remark 4 . If \mathring{\omega}\in L^{p} for some p , then \omega\in C([0,\infty);L^{q}_{\text{loc}}) for q=\min\{p,2\mu\} . Thus, for p\geq 2 , \omega is a renormalized solution in the sense of Diperna-Lions [ 18 ] . Furthermore, the assumption \mathring{\omega}\in L^{1} can be relaxed to \mathring{\omega} being a bounded measure, via mollification (see Section 8 of [ 41 ] and Corollary 5.11 ). Self-similar spiral solutions without rotational symmetry were also constructed for the generalized SQG equations by García and Gómez-Serrano [ 25 ] , where \left|\begin{array}[]{l}\partial_{t}\omega+u\cdot\nabla\omega=0,\\ u=-\nabla^{\perp}(-\Delta)^{-1+\frac{\gamma}{2}}\omega.\end{array}\right. To ensure the integrability of the nonlocal operator (-\Delta)^{-1+\frac{\gamma}{2}}\omega , they require \frac{1}{2}<\mu<\frac{1}{2-\gamma} . This range excludes the Euler case \gamma=0 . In contrast, our focus on the regime \mu>1 for the 2-D Euler explores a different region of the parameters. Extending our approach to the surface quasi-geostrophic (SQG) equations would be an interesting direction for future work. 1.2. Non-uniqueness in incompressible fluid equations The non-uniqueness of weak solutions of the Euler equations ( 1.1 ) has been an active field of study. The first explicit examples of non-unique solutions were constructed by Scheffer and Shnirelman [ 39 , 44 ] . A decisive breakthrough came with the