Abstract settings for tangential boundary stabilization of Navier--Stokes equations by high- and low-gain feedback controllers. (Q2490770)

The paper continues the analysis in [\textit{V. Barbu} \textit{et al}., Tangential boundary stabilization of Navier-Stokes equations, Mem. Am. Math. Soc, 852 (2006; Zbl 1098.35025)] on tangential boundary stabilization of Navier-Stokes (N-S) equations, \(d = 2, 3\), as deduced from well-posedness and stability properties of the corresponding linearized equations. Namely, having found a definite Riccati-based solution of the stabilization problem in question, the authors investigate what axiomatic properties, and in what topological setting depending on the dimension \(d = 2,3\), need to be satisfied by the linearized N-S equations, in order to guarantee local exponential stabilization of the nonlinear N-S equations, in the vicinity of a steady-state solution, by means of the same boundary feedback mechanism which has proved successful in the linearized dynamics. The work intends to complement the work cited above on two levels. The first level consists of extracting the key relevant ingredients occurring in the special Riccati-based solution presented in the previous work, \(d = 2,3\). These are then elevated into an abstract setting, capable then to imply and guarantee local, tangential feedback stabilization of the equilibrium solution for the full nonlinear N-S problem. This strategy results into two abstract frameworks based exclusively on the linearized N-S equations: more precisely, its semigroup well-posedness and its stability properties, linking a precise pair of {state, output}-spaces, which depends on \(d = 2,3\). The results may be applicable to other stabilizing boundary feedback operators, as well as to other parabolic-like equations of fluid dynamics. The second level intends to focus on studying the case \(d = 2\) directly (rather than as a treatment derived or implied by the analysis of the case \(d = 3\), as essentially done in the paper mentioned above). This way, one does not only re-obtain the Riccati-based, high-gain tangential local feedback stabilization results previously obtained by the authors, \(d = 2, 3\). But, in addition, it is possible to obtain local stabilization results for the N-S equations by different (spectral-based and Riccati-based) low-gain (tangential) feedback operators, in the case \(d = 2\). As a main result of the paper, it should be noted that a large class of linear feedback controllers that exponentially stabilize the (feedback) solutions of the N-S equations, \(d = 2, 3\), in a vicinity of an unstable stationary (equilibrium) solution is constructed. The corresponding stabilizing feedback controllers are either Riccati based, or else spectral based. In the first case, they are derived from an optimal control problem and related algebraic Riccati equations. In the second case, they are derived from the finite-dimensional projection onto the unstable subspace of the linearized N-S problem yielding a finite-dimensional controller via finite-dimensional control theory. It is important that the finite-dimensional stabilizing feedback controllers might arise from a numerical approximating scheme.