Deep Surrogate for Direct Time Fluid Dynamics
This work addresses the need for efficient fluid simulation surrogates in scientific and engineering domains, but it appears incremental as it builds on existing deep learning approaches for CFD.
The authors tackled the problem of simulating fluid dynamics for scientific and engineering applications by developing a novel direct time Graph Neural Network (GNN) architecture for irregular meshes, achieving small generalization errors and mitigating error accumulation on the Von Kármán's vortex street benchmark.
The ubiquity of fluids in the physical world explains the need to accurately simulate their dynamics for many scientific and engineering applications. Traditionally, well established but resource intensive CFD solvers provide such simulations. The recent years have seen a surge of deep learning surrogate models substituting these solvers to alleviate the simulation process. Some approaches to build data-driven surrogates mimic the solver iterative process. They infer the next state of the fluid given its previous one. Others directly infer the state from time input. Approaches also differ in their management of the spatial information. Graph Neural Networks (GNN) can address the specificity of the irregular meshes commonly used in CFD simulations. In this article, we present our ongoing work to design a novel direct time GNN architecture for irregular meshes. It consists of a succession of graphs of increasing size connected by spline convolutions. We test our architecture on the Von K{á}rm{á}n's vortex street benchmark. It achieves small generalization errors while mitigating error accumulation along the trajectory.