Physical AI exploration by a CAE engineer: MeshGraphNet, FNO, PyG experiments. Companion code for the "AI Demystification" series.
-
Updated
Apr 6, 2026 - Python
#
meshgraphnet
Here are 6 public repositories matching this topic...
Graph neural network surrogates for crash simulation: autoregressive deformation prediction with MeshGraphNets and MGN-T, from raw FE data to evaluation, in a lean PyTorch framework. Evaluated on deforming_plate and a bumper beam crash.
deep-learning cae structural-mechanics finite-element-method graph-neural-networks gnn scientific-machine-learning surrogate-model crash-simulation learned-simulation meshgraphnet meshgraphnets physics-ml crashworthiness mgn-t
-
Updated
Nov 12, 2026 - Python
Casting solidification GNN pipeline: STEP to healed CAD, Gmsh mesh, predicted solidification time, hot spots and shrinkage screening. The model is on Hugging Face (eugenmik/castsolid-gnn), Apache-2.0.
casting foundry solidification graph-neural-networks pytorch-geometric scientific-machine-learning surrogate-model fenicsx meshgraphnet
-
Updated
Oct 10, 2026 - Python
MeshGraphNet-style graph neural network (PyTorch) that predicts stress fields, natural frequency and fatigue life of FE bracket models
-
Updated
Nov 6, 2026 - Python
CFD-trained MeshGraphNet engineering portfolio for offline 3D heat-sink temperature-field prediction.
vtk heat-transfer computational-fluid-dynamics thermal-analysis pytorch-geometric graph-neural-network surrogate-model pyvista meshgraphnet engineering-ai
-
Updated
Jul 31, 2026
Physics-informed MeshGraphNet surrogate for operational storm surge forecasting
deep-learning pytorch noaa storm-surge graph-neural-network surrogate-model meshgraphnet ocean-modeling
-
Updated
Jul 1, 2026 - Python
Add this topic to your repo
To associate your repository with the meshgraphnet topic, visit your repo's landing page and select "manage topics."