Domain-decoupled physics-informed neural networks for inverse material property estimation from acoustic fields
JASA Express Letters · 01.11.2026 · 6 (11)
DOI: 10.1121/10.0047523 ↗Kurz gefasst
Fachartikel in der Zeitschrift JASA Express Letters, online erschienen am 16.11.2026. Verfasst von Paul Brand und 2 weiteren Autor:innen. Beteiligte Einrichtung: Fachgebiet Systemzuverlässigkeit, Adaptronik und Maschinenakustik, Technische Universität Darmstadt, Darmstadt, Hessen 64287, Germany. Die Arbeit nennt 13 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Technische Universität Darmstadt.
Abstract (Originalsprache)
Estimating material properties like Young's modulus and internal damping from radiated acoustic fields is challenging under sparse, noisy measurement conditions. This work employs sequential physics-informed neural networks with Fourier features to estimate these material properties for a vibrating plate. The inverse problem is solved by decoupling fluid and structural domains to effectively filter boundary differentiation and data noise. Multi-frequency statistical ensembles are introduced to quantify the uncertainty coming from neural network initialization robustness and learning biases. Following a validation of the method in a proof of concept on vast, low-noise data conditions, its limits under sparse, noisy data are demonstrated.
Autor:innen und Institutionen
- Fachgebiet Systemzuverlässigkeit, Adaptronik und Maschinenakustik, Technische Universität Darmstadt, Darmstadt, Hessen 64287, Germany.
ORCID-Profile: Paul Brand, Robert Feldmann, Tobias Melz
Förderung
- Technische Universität Darmstadt
Zitieren
APA
Brand, P., Feldmann, R., & Melz, T. (2026). Domain-decoupled physics-informed neural networks for inverse material property estimation from acoustic fields. JASA Express Letters, 6(11). https://doi.org/10.1121/10.0047523
Vancouver
Brand P, Feldmann R, Melz T. Domain-decoupled physics-informed neural networks for inverse material property estimation from acoustic fields. JASA Express Letters. 2026;6(11). doi:10.1121/10.0047523
BibTeX
@article{Brand2026,
author = {Brand, Paul and Feldmann, Robert and Melz, Tobias},
title = {Domain-decoupled physics-informed neural networks for inverse material property estimation from acoustic fields},
journal = {JASA Express Letters},
year = {2026},
volume = {6},
number = {11},
doi = {10.1121/10.0047523}
}
EUVIMED