This thesis focuses on learning a measuring method for the determination of the acoustic characteristics of porous materials within a reverberant room setting. Using a supervised learning approach, this research aims to estimate from sound field measurements in the industry standard Alpha Cabin, the most important material parameter that porous materials exhibit, namely the flow resistivity. The Alpha Cabin serves as a controlled test space that closely emulates the interior of a compact MPV-class vehicle. The central objective of the study is to design a convolutional neural network, that can be used to directly predict flow resistivity based on sound field measurements of the Alpha Cabin, instead of a more complicated approach using iterative optimization and repeated numerical simulations. For this, the strongly irregular sound field dominated by modes must be accurately modeled. The study considers various parameters, including frequency, humidity, porosity of the material, thickness of the absorber material, and flow resistivity, in order to generate a diverse dataset. This dataset serves as the basis for training the convolutional neural network to learn the underlying physical model, enabling it to generate results for unseen data.

Melden Sie sich hier an um Einladungen zu den Kolloquium-Vorträgen per E-Mail zu erhalten.
Register here to receive the invitations to colloquium talks via e-mail.

Zoom-Meeting-ID: 954 4073 3814
Passwort: 450783