A sparse formulation of the acoustic scattering of finite inhomogeneous structures

The objective of this fundamental research project is to derive a novel formulation for the acoustic scattering of finite structures with arbitrary shape and boundary conditions, whereby superior resolution and accuracy can be achieved with sparse representations and deep learning methodologies.

There are no formulations capable of accurately and efficiently predicting the scattered field under these challenging conditions, since all reported methods are either too simple to model complex structures, or too specific to allow for a natural generalization.

A formulation recently introduced by the applicant allows for the sparse recovery of acoustic wavefronts in rooms with arbitrary contents and properties, demonstrating accurate results with spatial resolution expanded by as much as 5 times.

The proposed research project advances the frontier of modelling acoustic scattering phenomena by including inhomogeneous edge diffraction effects commonly caused by complex object shapes and boundary conditions.The methods derived in this project allow for studying acoustic scattering phenomena such as the interaction between turbulent flow and wind turbine noise, as well as the propagation of noise in complex urban environments.

Besides contributing to the counteraction of noise in society, a deepened understanding of scattering phenomena enables the optimization of component size and weight in vehicles, which in turn translates into reduced energy consumption and a greener environment.