Laser ultrasonics and neural networks for the characterization of thin isotropic plates

Nondestructive testing has been performed on a thin copper plate. Lamb waves were generated and studied with a noncontact method, using a laser ultrasonic setup. A two-dimensional Fourier transform technique has been employed to obtain the dispersion curves, revealing numerous symmetric and antisymmetric modes. The inverse problem, in other words, the determination of the thickness and the elastic constants of the tested plate, has been solved by means of a feed-forward neural network. These parameters were then evaluated simultaneously, the dispersion curves being entirely fitted. The experimental results show good agreement with the theoretical model. This inversion method was found to be prompt and easy to automate.