2.0.0 Release

2.0.0 release of MGN (MeshGraphNets)

This code base contains PyTorch implementations of graph neural networks for CFD simulation surrogate development. The plan is to apply this code to predict volume fraction fields associated with CFD simulations of a representative column model (Fu et al., 2020) for solvent-based carbon capture within the CCSI2 project.

Pfaff et al., 2021. "Learning Mesh-Based Simulation with Graph Networks." International Conference on Learning Representations (ICLR) (2021)

Fu et al., 2020. "Investigation of Countercurrent Flow Profile and Liquid Holdup in Random Packed Column with Local CFD Data." Chemical Engineering Science 221 (2020)

What's Changed

• Adding correction for multigraph aspect by @Forbu in #6
• Add V2 code. by @pn51 in #10

New Contributors

• @Forbu made their first contribution in #6

Full Changelog: 1.0.0...2.0.0

Initial 1.0.0 Release

Initial 1.0.0 release of MGN (MeshGraphNets)

This code base contains PyTorch implementations of graph neural networks for CFD simulation surrogate development. The plan is to apply this code to predict volume fraction fields associated with CFD simulations of a representative column model (Fu et al., 2020) for solvent-based carbon capture within the CCSI2 project.

Pfaff et al., 2021. "Learning Mesh-Based Simulation with Graph Networks." International Conference on Learning Representations (ICLR) (2021)

Fu et al., 2020. "Investigation of Countercurrent Flow Profile and Liquid Holdup in Random Packed Column with Local CFD Data." Chemical Engineering Science 221 (2020)