Attractors for 2D-Navier-Stokes models with delays (Q704218)

The authors consider the Navier-Stokes equations with delay in 2D. The equation in question is \[ \begin{aligned} \partial_tu= \nu\Delta u-\nabla p- (u\nabla)u+ f- g(t, u_t)\quad &\text{in }(\tau,\infty)\times \Omega,\\ \text{div}(u)= 0,\;u= 0\quad &\text{on }(\tau,\infty)\times \partial\Omega,\;u(\tau, x)= u_0(x),\\ u(t,x)= \phi(t- \tau,x)\quad &\text{for }t\in (\tau- h,\tau),\;x\in\Omega.\end{aligned}\tag{1} \] Here, \(u_t(s)= u(t+ s)\), \(s\in (-h,0)\); the delay is encoded in \(\phi\), \(g\). \(\Omega\subset\mathbb{R}^2\) is a smooth bounded domain. In order to cast (1) into an abstract form, a standard \(L^2\)-setting is imposed on (1). E.g., \(H\) is the \(L^2\)-closure of \(V_0= \{u\in C^\infty_0(\Omega)\), \(\text{div}(u)= 0\}\), while \(V\) is the closure of \(V_0\) with respect to the norm \(\|\nabla u\|\). In order to take care of the delay, spaces \(C_H\), \(C_V\), \(L^2_H\) are introduced, where \(C_H= C^0([- h,0], H)\) etc. The delay is taken into account by \(g(t,\xi)\) which is a mapping from \(\mathbb R\times C_H\) to \(L^2(\Omega)^2\). The function \(g(t,\xi)\) is subject to a series of assumption involving measurability and two kinds of Lipschitz conditions. System (1) is then put into the abstract form \[ \partial_t u+\nu Au+ Bu= f+ g(t, u_t),\;u(0)= u_0\tag{2} \] along conventional lines. The authors first recall a result obtained in an earlier paper, stating that (under some assumptions) (2) admits a unique global solution in some suitable function space. The authors then proceed to show that (2) admits a global attractor. In this context, the usual notion of attractor does not suffice; it has to be replaced by a new notion, i.e. by a family \(\{A(t),\,t\in\mathbb{R}\}\) called pullback attractor, related to a family \(U(t,s)\) of evolution operators associated with (2). After proving the existence of an absorbing set, the authors then show that under suitable assumptions on the data, a unique, uniformly bounded pullback attractor for (2) exists. The paper concludes with an application to the case where (2) is supplied by a forcing term.