Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 26, 2026Journal of Applied Mechanics

MeshODENet: A Graph-Informed Neural Ordinary Differential Equation Neural Network for Simulating Mesh-Based Physical Systems

View Full Paper
Ask AI
Bookmark
Share

Authors

KLKangzheng LiuLMLeixin Ma

Discussion

Loading...

Member takes

Overview

This framework enhances accuracy and speed in simulating mesh-based physical systems, suggesting improved modeling techniques.

Key Points

  • The aim is to develop a more effective and efficient method for simulating complex physical systems using mesh data.
  • Introduced MeshODENet combining GNNs and Neural Ordinary Differential Equations.
  • Applied the framework to various structural mechanics problems.
  • Compared performance against traditional numerical solvers.
  • MeshODENet showed significantly better long-term predictive accuracy.
  • Demonstrated greater stability compared to autoregressive models.
  • Achieved substantial computational speed-ups over traditional methods.

Cite This Study

Liu et al. (2026) studied this question.

synapsesocial.com/papers/69c4cd73fdc3bde448919bdbhttps://doi.org/10.1115/1.4071488
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Graph Neural Networks for Mesh Generation and Adaptation in Structural and Fluid Mechanics2024 · 13 citations
  2. 2Graph Neural Networks for Mesh Generation and Adaptation in Structural and Fluid Mechanics2024 · 3 citations
  3. 3An Operator-Consistent Graph Neural Network for Learning Diffusion Dynamics on Irregular Meshes2025
  4. 4Mesh-based GNN surrogates for time-independent PDEs2024 · 43 citations
  5. 5Physics-informed MeshGraphNets (PI-MGNs): Neural finite element solvers for non-stationary and nonlinear simulations on arbitrary meshes2024