Physics-informed neural network for elastic wave-mode separation
Abstract
Mode conversion in nonhomogeneous elastic media makes it challenging to interpret physical properties accurately. Decomposing these modes correctly is crucial across various scientific areas. Recent machine learning approaches have been proposed to address this problem, utilizing the Helmholtz decomposition technique. In this paper, we investigate the capabilities of a physics-informed neural network (PINN) in separating P and S modes by solving a scalar Poisson equation. This scalar formulation offers a dimensionally scalable reduction in computational cost compared to the traditional vector formulation. We verify the proposed method in both homogeneous and realistic nonhomogeneous elastic models as showcases. The obtained separated modes closely match those from conventional numerical techniques, while exhibiting reduced transverse wave leakage.
Fields of Study & Subjects
Model Reduction and Neural Networks
Neural Networks and Reservoir Computing
Ultrasonics and Acoustic Wave Propagation
Publication details
- Journal / source
- Physical review. E
- Publisher
- American Physical Society (APS)
- Volume
- 113
- Issue
- 2
- Pages
- 025503–025503
- Publication date
- 9 January 2026
- Language
- EN