No label defined (Q3979144)

[For the entire collection see Zbl 0732.00028.] In the modeling and control of large flexible structures, one is often required to numerically recover one or more material parameters given data measurements at various points. Although these structures are in general very complex, in many cases the essential features can be developed by considering a fixed Euler-Bernoulli beam which is assumed to have Kelvin-Voigt damping. A fully Sinc-Galerkin method is presented for the numerical recovery of the stiffness parameter \(EI\) and the damping parameter \(c_ DI\) in the state space model \[ \begin{aligned}{\mathcal L}(EI,c_ DI)u &=f(x,t),\quad 0<x<1, t>0 \\ u(0,t) &=u(1,t)=0,\quad t>0\\ {\partial u \over\partial x}(0,t) &={\partial u \over\partial x}(1,t)=0,\quad t>0\\ u(x,0) &={\partial u \over\partial t}(x,0)=0,\quad 0\leq x\leq 1\end{aligned} \] with \[ {\mathcal L}(EI,c_ DI)u\equiv{\partial^ 2u \over\partial t^ 2}+{\partial^ 2 \over \partial x^ 2}\left(EI(x){\partial^ 2u \over\partial x^ 2}+c_ DI(x){\partial^ 3u \over\partial x^ 2\partial t}\right), \] given measurements of the data at the points \(\{(x_ p,t_ q)\}^{q=1,\dots,n_ q}_{p=1,\dots,n_ p}\) in the domain \((0,1)\times\mathbb{R}^ +\). The fully Sinc-Galerkin method in space and time has many salient features due both to the properties of the basis functions and the manner in which the problem is discretized. Perhaps the most distinctive feature of the method is the exponential convergence rate when solving the corresponding forward problems. Furthermore, the judicious choice of a conformal map provides approximate solutions which are valid on the infinite time interval rather than only on a truncated time domain. Finally, the discrete system requires no numerical integrations to fill either the coefficient matrices or the right-hand side matrix. All three features prove to be advantageous when solving the forward problems and hence the inverse problem. In Section 2, the Sinc-Galerkin system for the forward problem is considered and implementation details are discussed. The forward results are then incorporated into the finite-dimensional minimization problems in Section 3 with the discussion centering around the construction of the Tikhonov functional. Numerical results are presented in Section 4 along with a brief outline of the \(L\)-curve technique for determining the regularization parameter \(\alpha\). Examples are given which demonstrate the recovery of the individual parameters \(EI\) and \(c_ DI\) as well as the simultaneous recovery of both parameters.