MeshGraphNets
A PyTorch implementation of "Learning Mesh-based Simulation with Graph Networks" (Pfaff et al., 2021) to build and train graph neural network surrogates for CFD simulation.
Bartoldson, BrianR.↗
Engineering topics
Publications and source records attributed to Ng, BrendaM..
A PyTorch implementation of "Learning Mesh-based Simulation with Graph Networks" (Pfaff et al., 2021) to build and train graph neural network surrogates for CFD simulation.
A framework for building and training neural network surrogates for CFD simulation, with a focus on quickly simulating volume fraction dynamics for carbon capture efficiency prediction. Augments the Deep Fluids (Kim et al., 2019) approach as described in our paper "Latent Space Simulation for Carbon Capture Design Optimization" (Bartoldson et al., 2021).