On the uniqueness and structural stability of Couette-Poiseuille flow in a channel for arbitrary values of the flux

1. Problem Statement and Context

The Core Problem:
The paper addresses the Leray problem in the context of the steady-state Navier-Stokes equations within a two-dimensional, infinite, straight channel ($S = \mathbb{R} \times (-1, 1)$). Specifically, it investigates the uniqueness and structural stability of parallel flows (Couette-Poiseuille flows) characterized by an arbitrary flux $\Phi$.

Background & Motivation:

• Historical Context: Previous results on the existence and uniqueness of solutions converging to Poiseuille flow at infinity (as $|x| \to \infty$) typically required the flux magnitude $|\Phi|$ to be "small" (Amick) or "large" (Ladyzhenskaya & Solonnikov).
• The Gap: It remained an open question whether, for arbitrary flux values, the classical parallel flow solution is the unique solution in appropriate function spaces, or if other bounded solutions (potentially exhibiting flow reversal) could exist.
• The Specific Question: Is the linearized operator around a generic Couette-Poiseuille solution invertible for arbitrary flux? If so, this implies local uniqueness and structural stability.

Mathematical Formulation:
The authors consider the boundary value problem for velocity $\mathbf{u}$ and pressure $p$:
$$\begin{cases} -\Delta \mathbf{u} + (\mathbf{u} \cdot \nabla)\mathbf{u} + \nabla p = 0 & \text{in } S, \\ \nabla \cdot \mathbf{u} = 0 & \text{in } S, \\ \mathbf{u} = \mathbf{u}_{\text{wall}} & \text{on } \partial S, \\ \int_{-1}^1 \mathbf{u}(0, y) \cdot \mathbf{e}_1 \, dy = \Phi. \end{cases}$$
The base parallel flow is $\mathbf{u}^*(x,y) = F(y)\mathbf{e}_1$, where $F(y) = 3Ay^2 + By + C$. The flux is $\Phi = 2(A+C)$. The authors analyze perturbations $\mathbf{v} = \mathbf{u} - \mathbf{u}^*$ and study the linearized operator $L$ acting on $\mathbf{v}$.

2. Methodology

The paper employs a rigorous functional analytic approach combined with the study of ordinary differential equations (ODEs) derived from the stream function formulation.

Key Methodological Steps:

1. Stream Function Formulation:
   The linearized problem is reformulated using a stream function $\psi$, leading to a fourth-order partial differential equation. By applying a partial Fourier transform in the $x$-direction, the PDE is reduced to a family of non-homogeneous Orr-Sommerfeld equations (ODEs) parameterized by the frequency $\xi$:
   $$\phi_\xi'''' - 2\xi^2 \phi_\xi'' + \xi^4 \phi_\xi - i\xi [F(y)(\phi_\xi'' - \xi^2 \phi_\xi) - 6A\phi_\xi] = h_\xi(y)$$
   with boundary conditions $\phi_\xi(\pm 1) = \phi_\xi'(\pm 1) = 0$.

2. The McLeod Approach (Core Innovation):
   The authors extend and refine a technique originally introduced by J.B. McLeod (previously unpublished/unnoticed in fluid mechanics literature).

   • Transformation: They introduce an auxiliary function $\sigma(y) = \phi_\xi(y) / F(y)$.
   • Crucial Assumption: They impose conditions on parameters $A, B, C$ such that $F(y) \geq 0$ on $[-1, 1]$ and $F(y) > 0$ on $(-1, 1)$. This condition excludes flow reversal.
   • Energy Estimates: By multiplying the transformed ODE by $\sigma'$ and integrating, they derive sharp a priori estimates. A key step involves proving that the boundary terms arising from integration by parts are non-positive, which allows them to control the norms of the solution.

3. Isomorphism Proof:
   Using the derived estimates, they prove that the linearized operator $L$ is a continuous bijection (isomorphism) between specific Sobolev-type spaces and $L^2$ spaces.

   • Global Setting: Spaces $H(S)$ involving global $H^2$ norms.
   • Local Setting: Spaces $X_m$ involving locally bounded Dirichlet integrals (supremum of norms over unit intervals).

4. Nonlinear Stability via Contraction Mapping:
   Once the linear operator is proven to be an isomorphism, the Contraction-Mapping Theorem is applied to the nonlinear perturbation problem. This establishes that for small external forces (or small perturbations), the solution is unique and depends continuously on the data.

5. Symmetry Analysis:
   For a subclass of solutions possessing specific symmetries (e.g., $x$-even/odd or $y$-even/odd), the authors prove global uniqueness (uniqueness "in the large"), removing the "smallness" assumption required in the general nonlinear case.

6. Counter-Example Construction:
   To demonstrate the necessity of the "no flow reversal" assumption, they construct a counter-example using spectral instability results (referencing Grenier, Guo, and Nguyen). They show that if flow reversal is allowed, the linearized operator loses injectivity (has a non-trivial kernel).

3. Key Contributions and Results

Main Theorems:

• Theorem 1 (Global Isomorphism): Under the assumption that the flow does not reverse ($F(y) \geq 0$), the linearized Couette-Poiseuille operator is an isomorphism from the space of divergence-free vector fields with $H^2$ regularity to $L^2$. This holds for arbitrary flux values.
• Theorem 2 (Local Nonlinear Uniqueness): Consequently, the nonlinear problem admits a unique solution in a small neighborhood of the parallel flow for arbitrary flux, provided the perturbation is small in the $H^1$ norm.
• Theorem 3 (Local Isomorphism in $X_m$): The isomorphism property holds even in the "local" setting (spaces defined by locally bounded Dirichlet integrals), which is crucial for handling infinite domains without requiring global decay rates that might be too restrictive.
• Theorem 4 & Corollary 2: Structural stability is established in the local setting; small external forces yield unique solutions close to the parallel flow.
• Theorem 5 (Global Uniqueness for Symmetric Flows): If the base flow and the perturbation satisfy specific symmetry conditions (e.g., $B=0$ and symmetric boundary conditions), the parallel flow is the unique solution in the entire class of symmetric functions, regardless of the size of the perturbation (no "smallness" assumption needed).
• Theorem 6 (Necessity of Assumptions): If the condition preventing flow reversal is violated, the linearized operator is not invertible. There exist parameters where the operator has a non-trivial null space, implying potential bifurcation to non-parallel solutions.

4. Significance and Impact

1. Resolution of the Flux Restriction: The paper definitively removes the long-standing restriction on the magnitude of the flux $\Phi$. It proves that for Couette-Poiseuille flows in 2D channels, the parallel solution is locally unique and structurally stable for any flux value, provided the flow direction is maintained (no reversal).
2. Revival of McLeod's Work: The authors bring to light a powerful mathematical technique by J.B. McLeod that was previously unknown to the mathematical fluid mechanics community. This technique provides a robust method for proving invertibility of linearized operators in hydrodynamic stability.
3. Structural Stability: The results confirm that the classical parallel flow is a robust attractor for the Navier-Stokes equations in this geometry, provided the flow does not reverse.
4. Symmetry and Global Uniqueness: The paper provides a rare example of "global" uniqueness (without smallness assumptions) for the Navier-Stokes equations in an unbounded domain, achieved by exploiting symmetry properties.
5. Clarification of Flow Reversal: The counter-example highlights that flow reversal is the critical mechanism that can lead to non-uniqueness and instability, distinguishing the behavior of Couette-Poiseuille flows from purely Poiseuille flows in certain regimes.

In summary, this work provides a comprehensive and rigorous answer to the uniqueness of parallel flows in infinite channels, bridging a gap between local and global analysis and establishing that the "size" of the flux is not an obstacle to uniqueness, provided the flow remains unidirectional.