Nonlinear asymptotic stability of 2D Taylor-Couette flow in the exterior disk
Te Li, Ping Zhang, Yibin Zhang · 2025
Read Nonlinear asymptotic stability of 2D Taylor-Couette flow in the exterior disk by Te Li, Ping Zhang, Yibin Zhang (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-2503-20253.txt.
Introduction
Te Li, Ping Zhang, and Yibin Zhang wrote this 2025 math paper on the steadiness of a swirling flow between spinning walls, spread over the region outside a disk. The flow, called Taylor-Couette flow, fades at far range, which makes its decay hard to pin down. Past work showed that the flow cannot decay at the neat fast rate, and that at best a slower, mixed rate holds.
Old tools for such bounds assumed the decay shape stays fixed over time, which fails here. The team builds new estimates that let the decay shape move with time and ride on the flow's own wave parts. With those tools, they prove the flow stays stable under small nudges. They also recover the same threshold bound, one half, that is known for the flow in a bounded ring.
Catalog record
| CREATOR | Te Li, Ping Zhang, Yibin Zhang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 152 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-20253
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Nonlinear asymptotic stability of 2D Taylor-Couette flow in the exterior disk Te Li, Ping Zhang, Yibin Zhang Abstract . In this paper, we consider the asymptotic stability of the 2D Taylor-Couette flow in the exterior disk, with a small kinematic viscosity ν ≪ 1 \nu\ll 1 and a large rotation coefficient | B | |B| . Due to the degeneracy of the Taylor-Couette flow at infinity, we cannot expect the solution to decay exponentially in a space-time decoupled manner. As stated in previous work [ 20 ] , even space-time coupled exponential decay can not be expected, and at most, we can obtain space-time coupled polynomial decay. To handle the space-time coupled decay multiplier, the previous time-independent resolvent estimate methods no longer work. Therefore, this paper introduces time-dependent resolvent estimates to deal with the space-time coupled decay multiplier Λ k \Lambda_{k} . We remark that the choice of Λ k \Lambda_{k} is not unique, here we just provide one way to construct it. Finally, as an application, we derive a transition threshold bound of 1 2 \frac{1}{2} , which is the same as that for the Taylor-Couette flow in the bounded region. 1. Introduction We consider 2D incompressible Navier-Stokes equations in the exterior disk \Omega\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}\bigl\{x\in\mathbb{R}^{2}:\ |x|\geq 1\bigr\} : (1.1) \displaystyle\left\{\begin{aligned} &\partial_{t}V-\nu\Delta V+V\cdot\nabla V+\nabla P=0,\quad(t,x)\in\mathbb{R}^{+}\times\Omega,\\ &\text{div }V=0,\\ &V(0,x)=V_{0}(x),\end{aligned}\right. where V\in\mathbb{R}^{2} denotes the fluid velocity and P the scalar pressure function, which guarantees the divergence free condition of the velocity field, \nu designates the kinematic viscosity of the fluid. In this paper, we study the stability of a two-dimensional rotating flow in the exterior region \Omega , which is a steady solution to ( 1.1 ). The corresponding velocity, vorticity, and scalar pressure are given as follows: \displaystyle V^{\star}(r,\theta)=\left(\begin{array}[]{cc}-\sin\theta\\ \cos\theta\end{array}\right)\Big(Ar+\frac{B}{r}\Big){\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}\left(\begin{array}[]{cc}v^{\star,1}\\ v^{\star,2}\end{array}\right)},\quad W^{\star}(r){\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}\partial_{1}v^{\star,2}-\partial_{2}v^{\star,1}=2A, (1.2e) \displaystyle P^{\star}(r,\theta)=P^{\star}(r)\quad\text{with}\quad\partial_{r}P^{\star}(r)=\frac{1}{r}\Big(Ar+\frac{B}{r}\Big)^{2}, where A,B are two constants and we assume that \nu\ll|B| throughout this paper. The steady flow of the form ( 1.2 ) is well-known as Taylor-Couette flow. As a classic example of rotating flow, Taylor-Couette (TC) flow has been extensively studied in both the physical and mathematical contexts. In physical experiments, TC flow always refers to the fluid flow between two concentric cylinders, where the inner cylinder rotates while the outer cylinder can be either stationary or rotating. This flow configuration was first studied by G. I. Taylor in the 1920s [ 25 ] . Despite its simple structure, TC flow exhibits complex and rich stability properties, and has become a classic problem in fluid dynamics due to its rich flow patterns, as referenced in such review books as [ 12 , 10 , 23 , 24 , 18 ] . These corresponding flow patterns are controlled by the Reynolds number and the radius ratio of the cylinders. For example, the TC apparatus is also an excellent prototype for studying the transport properties of most astrophysical or geophysical rotating shear flows [ 14 ] . Depending on the rotational speed of each cylinder, different flow states can be achieved, corresponding to increasing or decreasing angular velocity and angular momentum. Likewise, regarding the issue of asymptotic stability at high Reynolds numbers [ 6 ] , TC flow, as a typical rotating shear flow, also exhibits rich decay properties. This paper primarily investigates the fully nonlinear long-time asymptotic stability behavior near the TC flow in the exterior disk. Earlier, the well-posedness and stability of the two-dimensional stationary Navier-Stokes equations in exterior regions [ 5 , 16 ] were widely studied and discussed. The following results are related to the fully nonlinear stability of the time-dependent Navier-Stokes equations near the 2D TC flow. • Under the condition |B|\ll 1 and with the domain being the exterior disk, [ 21 , 22 ] used perturbation theory to study the L^{p}-L^{q} estimate for the semigroup generated by the linearized operator, and further addressed the global well-posedness and nonlinear asymptotic stability for small initial data in the L^{2} space. • In [ 2 ] , nonlinear enhanced dissipation and transition thresholds near the TC flow were obtained in the bounded annular region (r,\theta)\in[1,R]\times\mathbb{T} . Since the domain is compact, the shear effect of the TC flow is non-degenerate, i.e., (A+\frac{B}{r^{2}})^{\prime} is non-vanishing. In the case when \nu is sufficiently small and |B| is large enough, that is, when the Reynolds number is high and the rotational effects are strong, the fully nonlinear long-time asymptotic stability problem of the 2D TC flow in the exterior disk is both interesting and challenging. 1.1. Derivation of the perturbation equation Let (V^{\star},P^{\star}) be determined by ( 1.2 ). We denote U{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}V-V^{\star} , p{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}P-P^{\star} , and consider the following perturbation equations of the system ( 1.1 ) near the 2D TC flow: (1.3) \displaystyle\left\{\begin{aligned} &\partial_{t}U-\nu\Delta U+V^{\star}\cdot\nabla U+U\cdot\nabla V^{\star}+U\cdot\nabla U+\nabla p=0,\\ &\mathop{\rm div}\nolimits U=0,\quad(t,x)\in\mathbb{R}^{+}\times\Omega,\\ &U(t=0)=U(0).\end{aligned}\right. We shall present a local well-posedness result concerning the system ( 1.3 ) in Appendix A . Since this paper aims at investigating the impact of degeneracy, which is caused by the external region, on the nonlinear stability of TC flow, we impose the impermeable and Navier-slip boundary conditions for the velocity field: (1.4) U\cdot\mathbf{n}|_{\partial\Omega}=0,\quad\bigl(\partial_{1}U^{2}-\partial_{2}U^{1}\bigr)|_{\partial\Omega}=0. Let W\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}\mathop{\rm curl}\nolimits U{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}\partial_{1}U^{2}-\partial_{2}U^{1} be the vorticity of U , and we rewrite ( 1.3 - 1.4 ) in the equivalent vorticity-stream formulation: (1.5) \displaystyle\left\{\begin{aligned} &\partial_{t}W-\nu\Delta W+V^{\star}\cdot\nabla W+U\cdot\nabla W=0,\qquad(t,x)\in\mathbb{R}^{+}\times\Omega,\\ &\Delta\varPhi=W,\quad U=\nabla^{\perp}\varPhi{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}\begin{pmatrix}-\partial_{2}\varPhi,\partial_{1}\varPhi\end{pmatrix}^{T},\\ &W|_{\partial\Omega}=\varPhi|_{\partial\Omega}=0,\quad W(t=0)=W(0).\end{aligned}\right. Let \mathbf{e_{r}}=(\cos\theta,\sin\theta) and \mathbf{e_{\theta}}=(-\sin\theta,\cos\theta). In polar coordinates, ( 1.5 ) is equivalent to (1.6) \displaystyle\left\{\begin{aligned} &\partial_{t}W-\nu\Big(\partial_{r}^{2}+\frac{1}{r}\partial_{r}+\frac{1}{r^{2}}\partial_{\theta}^{2}\Big)W+\Big(A+\frac{B}{r^{2}}\Big)\partial_{\theta}W+\frac{1}{r}(\partial_{r}\varPhi\partial_{\theta}W-\partial_{\theta}\varPhi\partial_{r}W)=0,\\ &\Big(\partial_{r}^{2}+\frac{1}{r}\partial_{r}+\frac{1}{r^{2}}\partial_{\theta}^{2}\Big)\varPhi=W,\quad(t,r,\theta)\in\mathbb{R}^{+}\times[1,\infty)\times\mathbb{T},\\ &U=U^{r}\mathbf{e_{r}}+U^{\theta}\mathbf{e_{\theta}}\quad\hbox{with}\quad U^{r}=-\frac{\partial_{\theta}\varPhi}{r}\quad\hbox{and}\quad U^{\theta}=\partial_{r}\varPhi,\\ &W(t=0)=W(0).\end{aligned}\right. By expanding the functions (W,\varPhi,U^{r},U^{\theta}) in Fourier series, we write (1.7) \begin{split}&W(t,r,\theta)=\sum\limits_{k\in\mathbb{Z}}e^{ik\theta}\widehat{W}_{k}(t,r),\quad\varPhi(t,r,\theta)=\sum\limits_{k\in\mathbb{Z}}e^{ik\theta}\widehat{\varPhi}_{k}(t,r),\quad\\ &U^{r}(t,r,\theta)=\sum\limits_{k\in\mathbb{Z}}e^{ik\theta}\widehat{U}^{r}_{k}(t,r),\quad U^{\theta}(t,r,\theta)=\sum\limits_{k\in\mathbb{Z}}e^{ik\theta}\widehat{U}^{\theta}_{k}(t,r).\end{split} Then by substituting ( 1.7 ) into ( 1.6 ) and comparing the Fourier coefficients of the resulting equations, we obtain (1.8) \displaystyle\left\{\begin{aligned} &\partial_{t}\widehat{W}_{k}-\nu\Big(\partial_{r}^{2}+\frac{1}{r}\partial_{r}-\frac{k^{2}}{r^{2}}\Big)\widehat{W}_{k}+\Big(A+\frac{B}{r^{2}}\Big)ik\widehat{W}_{k}\\ &\qquad+\frac{i}{r}\sum_{\ell\in\mathbb{Z}}\big((k-\ell)\partial_{r}\widehat{\varPhi}_{\ell}\widehat{W}_{k-\ell}-\ell\widehat{\varPhi}_{\ell}\partial_{r}\widehat{W}_{k-\ell}\big)=0,\quad(t,r,\theta)\in\mathbb{R}^{+}\times[1,\infty)\times\mathbb{T}\\ &\Big(\partial_{r}^{2}+\frac{1}{r}\partial_{r}-\frac{k^{2}}{r^{2}}\Big)\widehat{\varPhi}_{k}=\widehat{W}_{k},\quad\widehat{U}^{r}_{k}=-\frac{ik}{r}\widehat{\varPhi}_{k},\quad\widehat{U}^{\theta}_{k}=\partial_{r}\widehat{\varPhi}_{k},\\ &\widehat{W}_{k}(t=0)=\widehat{W}_{k}(0).\end{aligned}\right. To get rid of the terms involving \frac{1}{r}\partial_{r} and A in ( 1.8 ), we introduce (1.9) \displaystyle w_{k}{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}r^{\frac{1}{2}}e^{ikAt}\widehat{W}_{k},\quad\varphi_{k}{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}r^{\frac{1}{2}}e^{ikAt}\widehat{\varPhi}_{k},\quad u^{r}_{k}{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}r^{\frac{1}{2}}e^{ikAt}\widehat{U}^{r}_{k},\quad u^{\theta}_{k}{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}r^{\frac{1}{2}}e^{ikAt}\widehat{U}^{\theta}_{k}. It is easy to observe that \displaystyle\|W\|_{L^{2}(\Omega)}^{2}\approx\sum_{k\in\mathbb{Z}}\|\widehat{W}_{k}\|_{L^{2}((1,\infty);rdr)}^{2}=\sum_{k\in\mathbb{Z}}\|w_{k}\|_{L^{2}((1,\infty);dr)}^{2}, \displaystyle\|\partial_{r}W\|_{L^{2}(\Omega)}^{2}\approx\sum_{k\in\mathbb{Z}}\|\partial_{r}\widehat{W}_{k}\|_{L^{2}((1,\infty);rdr)}^{2}=\sum_{k\in\mathbb{Z}}\|r^{\frac{1}{2}}\partial_{r}(r^{-\frac{1}{2}}w_{k})\|_{L^{2}((1,\infty);dr)}^{2}, which will be frequently used throughout this paper. Moreover, we deduce from ( 1.8 ) that (1.10a) \displaystyle\partial_{t}w_{0}+\mathcal{T}_{0}w_{0}=r^{-\frac{1}{2}}\sum_{\ell\in\mathbb{Z}\backslash\{0\}}i\ell\partial_{r}\big(r^{-1}\varphi_{\ell}w_{-\ell}\big),\quad(t,r)\in\mathbb{R}^{+}\times[1,+\infty), (1.10b) \displaystyle\partial_{t}w_{k}+\mathcal{T}_{k}w_{k}=\sum_{\ell\in\mathbb{Z}\backslash\{0\}}i\bigl[r^{-\frac{1}{2}}\ell\partial_{r}\big(r^{-1}\varphi_{\ell}w_{k-\ell}\big)-kr^{-1}\partial_{r}(r^{-\frac{1}{2}}\varphi_{\ell})w_{k-\ell}\bigr]-ikr^{-\frac{3}{2}}u^{\theta}_{0}w_{k}, (1.10c) \displaystyle\Big(\partial_{r}^{2}-\frac{k^{2}-\frac{1}{4}}{r^{2}}\Big)\varphi_{k}=w_{k}, (1.10d) \displaystyle u^{r}_{k}=-i\frac{k}{r}\varphi_{k},\quad u^{\theta}_{k}=r^{\frac{1}{2}}\partial_{r}(r^{-\frac{1}{2}}\varphi_{k})=\partial_{r}\varphi_{k}-\frac{1}{2r}\varphi_{k}, (1.10e) \displaystyle w_{k}|_{r=1}=\varphi_{k}|_{r=1}=0,\qquad\textrm{for any}\ k\in\mathbb{Z}, where \mathcal{T}_{k} is defined as (1.11) \displaystyle\mathcal{T}_{k}{\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}-\nu\Big(\partial_{r}^{2}-\frac{k^{2}-\frac{1}{4}}{r^{2}}\Big)+i\frac{kB}{r^{2}},\quad D(\mathcal{T}_{k}){\mathrel{\mathop{\kern 0.0pt=}\limits^{\hbox{\footnotesize def}}}}H^{2}\bigl((1,\infty);dr\bigr)\cap H^{1}_{0}\bigl((1,\infty);dr\bigr). So far, we have transformed the original problem into exploring the dynamic and long-time behavior of the system ( 1.10 ). 1.2. Key ideas It’s well known that, for transport-diffusion equations, the enhanced dissipation scale is closely related to the degree of degeneracy of the velocity field, see for instance [ 1 , 4 ] . However, as pointed out in [ 20 ] , due to the infinite degeneracy of TC flow in the exterior disk (i.e., \partial_{r}^{n}(r^{-2}) vanishes at infinity for any n\geq 0 ), the previous results are no longer applicable to e^{-\mathcal{T}_{k}t} , which motivates us to investigate the mapping properties of e^{-\mathcal{T}_{k}t} in some space-time weighted spaces in [ 20 ] . In particular, we observe from Theorem 1.2 of [ 20 ] that due to the degeneracy at infinity, the non-homogeneous equation ( 1.10b ) cannot exhibit exponential decay. Specifically, Lemma 1.1 (Theorem 1.2, [ 20 ] ) . For any positive continuous functions a_{1}(r) , a_{2}(r) and strictly decreasing function