User: Guest  Login
Title:

Learning Lagrangian Fluid Mechanics with E(3)-Equivariant Graph Neural Networks

Document type:
Buchbeitrag
Author(s):
Toshev, Artur P.; Galletti, Gianluca; Brandstetter, Johannes; Adami, Stefan; Adams, Nikolaus A.
Abstract:
We contribute to the vastly growing field of machine learning for engineering systems by demonstrating that equivariant graph neural networks have the potential to learn more accurate dynamic-interaction models than their non-equivariant counterparts. We benchmark two well-studied fluid-flow systems, namely 3D decaying Taylor-Green vortex and 3D reverse Poiseuille flow, and evaluate the models based on different performance measures, such as kinetic energy or Sinkhorn distance. In addition, we i...     »
Keywords:
Equivariance; Fluid mechanics; Graph Neural Networks; Lagrangian Methods; Smoothed Particle Hydrodynamics
Dewey Decimal Classification:
620 Ingenieurwissenschaften
Book title:
Lecture Notes in Computer Science
Publisher:
Springer Nature Switzerland
Year:
2023
Pages:
332-341
Covered by:
Scopus
Print-ISBN:
97830313829879783031382994
Language:
en
DOI:
doi:10.1007/978-3-031-38299-4_35
WWW:
https://link.springer.com/chapter/10.1007/978-3-031-38299-4_35
TUM Institution:
Lehrstuhl für Aerodynamik und Strömungsmechanik
 BibTeX