No label defined (Q4805871)

The paper studies the stabilization of the Navier-Stokes equation in a domain \(\Omega\subset{\mathbb R}^d\), \(d=2\), or \(3\). The equation with state \(y(x,t)=(y_1,\dots,y_d)\) is described in weak form by \[ y_t(t)+\nu Ay(t)+By(t)=P(mu)+Pf_0,\quad y(0)=y_0, \] where \(\nu\) denotes the viscosity constant; \(A\) the Stokes operator defined by \((Ay,w)=\sum_{i=1}^d\int_\Omega\nabla y_i\cdot\nabla w_i dx\); \(B\) the operator such that \((By,w)=\sum_{i,j=1}^d\int_\Omega y_iD_iy_jw_j dx\); \(P\) the Leray projector associated with the unknown pressure \(p(x,t)\); and \(u(x,t)=(u_1,\dots,u_d)\) the control input which enters the equation through the characteristic function \(m(x)\) of the open subset \(\omega\subset\Omega\). Weak solutions \(y(t)\) are sought in the space \(L^\infty_{\text{loc}}({\mathbb R}^+; H)\cap L^2_{\text{loc}} ({\mathbb R}^+; V)\), where \(H=\{y\in(L^2(\Omega))^d\), \(\nabla\cdot y=0\) on \(\Omega\), \(y\cdot n=0\) on \(\partial\Omega\}\) and \(V=\{y\in (H^1_0(\Omega))^d\), \(\nabla\cdot y=0\}\). Let \((y_e,p_e)\) be a steady state to the equation. Let \(R\) be the solution to the algebraic Riccati equation associated with an optimal control problem for the linearized Navier-Stokes equation, and set \(u\) as the feedback: \(u=-mR(y-y_e)\). Then it is shown that, as long as the initial data \(y_0\) is close to \(y_e\), the state \(y(t)\) approaches to \(y_e\) exponentially as \(t\to\infty\).