Application of artificial neural networks to predict parameters describing porous samples.

What is it about?

Artificial neural networks are used to determine the Johnson-Champoux-Allard fluid-acoustic parameters characterizing porous samples. In particular, a hierarchical procedure, based on shallow network in series, is applied. The algorithm is trained with data easy to determine experimentally, namely, thickness and absorption coefficient. The results demonstrate that the procedure is able to provide robust predictions with limited error.

Featured Image

Why is it important?

The application of machine learning allows an accurate prediction of the Johnson-Champoux-Allard parameters starting from data easily measurable - the absorption coefficient and thickness - without requiring the setup of several experimental tools, experimental campaigns, testing rigs, samples, or the use of complex mathematical models.