Generalization capabilities of MeshGraphNets to unseen geometries for fluid dynamics

2024/08/12 by Robin Schmöcker, Schmöcker, Robin, Alexander Henkes +5 · 1 citation
Computer Science · #Graph Theory and Algorithms #Computational Physics and Python Applications #Data Management and Algorithms

paper · pdf · doi:10.48550/arxiv.2408.06101

Abstract

This works investigates the generalization capabilities of MeshGraphNets (MGN) [Pfaff et al. Learning Mesh-Based Simulation with Graph Networks. ICML 2021] to unseen geometries for fluid dynamics, e.g. predicting the flow around a new obstacle that was not part of the training data. For this purpose, we create a new benchmark dataset for data-driven computational fluid dynamics (CFD) which extends DeepMind's flow around a cylinder dataset by including different shapes and multiple objects. We then use this new dataset to extend the generalization experiments conducted by DeepMind on MGNs by testing how well an MGN can generalize to different shapes. In our numerical tests, we show that MGNs can sometimes generalize well to various shapes by training on a dataset of one obstacle shape and testing on a dataset of another obstacle shape.

Cited by

• Multiscale Graph Neural Network for Turbulent Flow-Thermal Prediction Around a Complex-Shaped Pin-Fin

Related

• Scientific Computing Algorithms to Learn Enhanced Scalable Surrogates for Mesh Physics
• MeshMask: Physics-Based Simulations with Masked Graph Neural Networks
• Finite Volume Features, Global Geometry Representations, and Residual Training for Deep Learning-based CFD Simulation
• MeshDQN: A Deep Reinforcement Learning Framework for Improving Meshes in Computational Fluid Dynamics
• Simulating Liquids with Graph Networks