Lagrangian controllability in perforated domains
Mitsuo Higaki, Jiajiang Liao, Franck Sueur · 2025
Read Lagrangian controllability in perforated domains by Mitsuo Higaki, Jiajiang Liao, Franck Sueur (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-2509-14913.txt.
Introduction
Mitsuo Higaki, Jiajiang Liao, and Franck Sueur ask whether a patch of fluid can be steered to a target by remote action inside a perforated domain, in a 2025 fluid mechanics paper. The domain is riddled with many small holes, like a porous solid. Lagrangian controllability of viscous flow near solid walls is notoriously hard because of boundary layers.
The paper treats perforated domains where holes occupy a small fraction of the volume. Results distinguish cases quantitatively by hole size, hole spacing, and the size of the initial flow data. The approach builds on recent homogenization results for time-dependent problems, which replace the perforated medium by an effective one, plus stability estimates for flows in measure. Classical Runge-type approximation arguments complete the proof strategy for the Navier-Stokes setting.
Catalog record
| CREATOR | Mitsuo Higaki, Jiajiang Liao, Franck Sueur |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-14913
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Lagrangian controllability in perforated domains Mitsuo Higaki, Jiajiang Liao, Franck Sueur Abstract The question at stake in Lagrangian controllability is whether one can move a patch of fluid particles to a target location by means of remote action in a given time interval. In the last two decades, positive results have been obtained both for the incompressible Euler and Navier-Stokes equations. However, for the latter, the case where the fluid is contained within domains bounded by solid boundaries with the no-slip condition has not been addressed, with respect to the difficulty caused by viscous boundary layers. In this paper, we investigate the Lagrangian controllability of viscous incompressible fluid in perforated domains for which the fraction of volume occupied by the holes is sufficiently small. Moreover, we quantitatively distinguish situations depending on the parameters for holes (diameter and distance) and for fluid (size of the initial data). Our approach relies on recent results on homogenization for evolutionary problems and on weak-strong stability estimates in measure of flows, alongside classical results on Runge-type approximations for elliptic equations and on Cauchy-Kowalevsky-type theorems for equations with analytic coefficients. Here, homogenization refers to the vanishing viscosity limit outside a porous medium, where (after scaling in time) the Navier-Stokes equations are homogenized to the Euler or Darcy equations. Indeed, in the proof, we act on the Navier-Stokes equations by strong and fast forcing to leverage inviscid approximations, which is a standard technique in the theory of controllability. Contents 1 Introduction A classical problem in the controllability of the partial differential equations (PDEs) is determining if it is feasible to navigate the state of a physical system described by a PDE to a desired or specified target state within an imparted time. For instance, if one considers the incompressible Euler or Navier-Stokes equations, the problem is to ask if the state of the fluid velocity field can be modified; see in particular [ 38 , 9 , 10 , 19 , 11 , 12 ] . On the other hand, a more recent problem under consideration is the Lagrangian controllability. This problem seeks to determine whether it is feasible to navigate a patch of fluid particles to a target location within an imparted time by means of a remote action, typically imposed on a portion of the boundaries surrounding the fluid. In these last two decades, positive results for the Lagrangian controllability have been obtained both for the incompressible Euler and Navier-Stokes equations; see [ 27 , 21 , 22 , 23 , 28 , 36 ] . Let us briefly summarize the known results for the Lagrangian controllability of the Navier-Stokes equations with macroscopic bodies with solid boundaries. For the case of Navier-slip conditions on the boundaries, substantial progress has been achieved in [ 36 ] ; see also [ 37 ] studying the MHD equations. However, for the case of the no-slip boundary condition, no results have been reported even in the analytic framework. A significant challenge arising from the no-slip boundary condition is the considerable difficulty in estimating the impact of viscous boundary layers near boundaries. The standard method for addressing controllability, as in [ 13 , 12 , 36 , 35 ] , involves time scaling to introduce small viscosity in the Navier-Stokes equations, thereby leveraging the controllability of the inviscid equations. The underlying idea is to act on the viscous equations by a strong and fast control to bring them closer to the inviscid counterparts. Therefore, the method requires studying the vanishing viscosity limit of the Navier-Stokes equations, a problem recognized as notably difficult when the no-slip conditions are imposed on macroscopic boundaries; see, for example, [ 39 ] . In this paper, we consider the Lagrangian controllability for non-analytic boundaries with the no-slip condition, focusing on the case when boundaries are small. In fact, the limit of vanishing viscosity has been studied outside a porous medium, that is, in a domain perforated by holes with no-slip condition whose proportion in the total fluid is sufficiently small relative to viscosity [ 29 , 34 , 26 ] . In these situations, the local Reynolds number is bounded and thus the viscous boundary layers are negligible, enabling us to perform a simple energy computation similar to that in [ 31 ] and to avoid the Prandtl boundary layers. A prototypical result [ 29 , Theorem 1] says that a Leray-Hopf weak solution of the Navier-Stokes equations with viscosity \nu converges to the solution to the Euler equations in the limit \nu\to 0 when the fluid domain is the exterior to a single body whose volume is sufficiently smaller than \nu . This result is extended by [ 34 ] quantitatively by introducing parameters for holes (diameter and distance) and for fluid (viscosity and size of the initial data), and by [ 26 ] to the critical and supercritical cases of parameters where the limit equations are shown to be the Euler-Brinkman and Darcy equations, respectively, contrasting with the Euler equations in the subcritical case. The quantitative estimates in [ 34 , 26 ] are pivotal when considering the Lagrangian controllability in a perforated domain. This paper is organized as follows. In Section 2 , we present a problem of the Lagrangian controllability in partially perforated domains, and prove the main result Theorem 2.1 . In Section 3 , we present a problem of the Lagrangian controllability in fully perforated domains, and prove the main result Theorem 3.2 . In Appendix A , we collect several notions from differential geometry. In Appendix B , we recall the notion of a flow associated with a Leray-type solution of the Navier-Stokes equations. The materials in Appendices A – B will be referenced throughout this paper. 2 Lagrangian controllability in partially perforated domains Let us consider the 3d Navier-Stokes equations with viscosity 1 and forcing F^{\varepsilon} \left\{\begin{array}[]{ll}\partial_{t}U^{\varepsilon}-\Delta U^{\varepsilon}+U^{\varepsilon}\cdot\nabla U^{\varepsilon}+\nabla P^{\varepsilon}=F^{\varepsilon}&\mbox{in}\ (0,T)\times\Omega^{\varepsilon},\\ \nabla\cdot U^{\varepsilon}=0&\mbox{in}\ (0,T)\times\Omega^{\varepsilon},\\ U^{\varepsilon}=0&\mbox{on}\ (0,T)\times\partial\Omega^{\varepsilon},\\ U^{\varepsilon}=U^{\varepsilon}_{0}&\mbox{on}\ \{0\}\times\Omega^{\varepsilon}.\end{array}\right. (2.1) Assuming \Omega^{\varepsilon} to be a partially perforated domain, we are interested in the following Lagrangian controllability problem: move fluid particles that are initially in a polluted zone P_{0}^{\varepsilon} in a set K toward a safe zone P_{1} without pollution away from K in some time T_{c}>0 . The way to move the particles is by remote action: regard the term F^{\varepsilon} in ( 2.1 ) as a control forcing moving P_{0}^{\varepsilon} to P_{1} and assume that F^{\varepsilon} is smooth and supported in a control zone \Omega_{c} . The control zone \Omega_{c} is supposed to be away from both K and P_{1} . An intrinsic feature of this problem is that one cannot hope to move all the fluid particles, not even the full Lebesgue measure of the patch of particles. This occurs because the particles close to the no-slip boundaries are unable to escape from their, yet shrinking, connected component. Hence, we will focus on the goal of moving most of the particles. The precise definitions of \Omega^{\varepsilon},P_{0}^{\varepsilon},P_{1} , and \Omega_{c} are as follows. In this paper, we regard the 3 -torus \mathbb{T}^{3} as the cube [0,1]^{3} with identified sides. It is equivalent to regard \mathbb{T}^{3} as the orbit space of \mathbb{R}^{3} under the action of translations by \mathbb{Z}^{3} . Then we can equip \mathbb{T}^{3} with a Riemannian