A phenomenological model for sonic crystals based on artificial neural networks

Abstract

[EN] A phenomenological model that simulates the acoustic attenuation behavior of sonic crystals is developed in this paper. The input of the model is a set of parameters that characterizes each experimental setup, and the output is a simulation of the associated attenuation spectrum. The model consists of a combination of a multiresolution analysis based on wavelet functions and a set of artificial neural networks. An optimized coupling of these tools allows us to drastically reduce the experimental data needed, and to obtain a fast computational model that can be used for technological purposes. © 2006 Acoustical Society of America.