A selection principle for 2D steady Euler flows via the vanishing viscosity limit

TL;DR

This paper characterizes the vanishing viscosity limits of 2D steady flows, showing they are restricted to flows with constant vorticity or specific shear flows, thus aiding in selecting physically relevant solutions.

Contribution

It provides a complete characterization of vanishing viscosity limits for 2D steady Euler flows in various domains, relaxing previous assumptions about streamlines.

Findings

Vanishing viscosity limits in bounded domains are flows with constant vorticity.

In infinite strips, limits are restricted to constant, Couette, or Poiseuille flows.

Chaotic streamlines are null with respect to Lebesgue measure, aiding analysis.

Abstract

The 2D Euler system, which governs inviscid incompressible fluid flow, can admit infinitely many steady solutions in a given domain with slip boundary conditions. To select physical classical solutions, we investigate the vanishing viscosity limits of the steady Navier-Stokes system. The vanishing viscosity limits in periodic strips or bounded connected domains are completely characterized, even when strong boundary layers may appear. More precisely, we show that the only vanishing viscosity limits in a bounded connected domain are flows with constant vorticity. The significance of this result is that the approximating Navier-Stokes solutions are not required to have nested closed streamlines, an essential assumption in the century-old Prandtl-Batchelor theorem. For flows in an infinitely long strip, if the viscous velocity (but not the pressure) is periodic in the strip direction, we show that the only vanishing viscosity limits are constant flows, Couette flows, and Poiseuille flows. The proof relies on a delicate analysis of the streamlines for both viscous and inviscid flows, in which a key observation is that the set of chaotic streamlines for the Euler flow is null with respect to two-dimensional Lebesgue measure. The second result depends not only on the first but also on a powerful rigidity theorem that any non-shear steady classical Euler flow in a periodic strip must have closed streamlines, established via an analysis of streamlines and a novel total curvature estimate.