ABSTRACT
An inverse problem is investigated for identifying physical parameters in flexible structures using the dynamic data taken in a non-destructive manner. As a flexible structure a uniform/nonuniform cantilevererd beam is considered whose flexural stiffness and both air and structural dampings are unknown. These physical parameters may be spatially constant or distributed (i.e., spatially invariant or varying). In any case the unknown profiles are parametrized by expanding them using basis functions and the inverse problem is converted to one of determining expansion coefficients. This parameter estimation problem is solved by minimizing the norm of distance between measured data and solutions of finite-dimensional model approximated by the Galerkin scheme from the original infinite-dimensional mathematical model described by an Euler-Bernoulli partial differential equation. Then the identification algorithm is derived based on the Gauss-Newtonian least-squares principle. Several numerical simulations and the result of experiment are presented to show the efficacy of the proposed algorithm.
