Direct and inverse problems for a phase-field model with memory (Q5945106): Difference between revisions

The author considers the problem of recovering the scalar memory kernel \(h:[0,T]\to {\mathbb R}\) in the phase field semilinear integrodifferential model \[ \alpha_{1,1}D_tu(t,x) + \alpha_{1,2}D_tv(t,x) \] \[ =A(x,D_x)u(t,x) + \int_0^t h(t-s)A(x,D_x)u(s,x)ds + f(u(t,x),v(t,x)) + f_1(t,x), \] \[ \alpha_{2,1}D_tu(t,x) + \alpha_{2,2}D_tv(t,x) \] \[ =B(x,D_x)u(t,x) + g(u(t,x),v(t,x)) + g_1(t,x), \quad (t,x)\in (0,T]\times \Omega. \] Here \(A(x,D_x)\) and \(B(x,D_x)\) are strongly elliptic linear differential operators and the constant matrix \(\{\alpha_{i,j}\}_{i,j=1}^2\) is invertible with positive elements on the principal diagonal. Moreover, \(f,g:{\mathbb R}^2\to {\mathbb R}\) and \(f_1,g_1:[0,T]\times \Omega\to {\mathbb R}\) are given functions. \noindent The functions \(u\) and \(v\), standing for the temperature and the phase field, must satisfy, in addition to the initial and boundary conditions \[ u(0,x)=u_0(x),\qquad v(0,x)=v_0(x),\quad x\in \Omega, \] \[ h_0(x)D_\nu u(t,x) + \int_0^t h(t-s)D_\nu u(s,x)ds =0,\qquad D_\nu v(t,x) = 0,\quad (t,x)\in (0,T]\times \Omega, \] also the additional information \[ \int_\Omega \phi(x)u(t,x)dx = b(t), \quad t\in [0,T]. \] The functions \(\phi,h_0:\Omega \to {\mathbb R}\) and \(b:[0,T]\to {\mathbb R}\) are given, while \(\nu\) denotes the outward normal to \(\partial \Omega\). \noindent The author associates with the previous problem a corresponding nonlinear identification problem in a Banach space which is solved by the joint use of semigroup-theory and fixed-point techniques. The abstract results are then applied to the specific phase field problem.