Design of acoustic absorbing metasurfaces using a data-driven approach

Abstract The design of acoustic metasurfaces with desirable properties is challenging due to their artificial nature and the large space physical geometrical parameters. Until recently, strategies were primarily based on numerical simulations finite-element or finite-difference time-domain methods, which are limited in terms computational speed complexity. Here, we present an efficient two-stage data-driven approach for analyzing designing membrane-type metasurface absorbers characteristics. In first stage, a forward model consisting neural network trained map input, comprising membrane parameters, observed sound absorption spectrum. second learned inverted infer input parameters that produce desired response. serve as network, estimated by minimizing loss function between profile output model. Two devices then fabricated using measured responses show very close agreement responses.