Combining Differentiable PDE Solvers and Graph Neural Networks for Fluid Flow Prediction

  • 2020-07-08 21:23:19
  • Filipe de Avila Belbute-Peres, Thomas D. Economon, J. Zico Kolter
  • 65

Abstract

Solving large complex partial differential equations (PDEs), such as thosethat arise in computational fluid dynamics (CFD), is a computationallyexpensive process. This has motivated the use of deep learning approaches toapproximate the PDE solutions, yet the simulation results predicted from theseapproaches typically do not generalize well to truly novel scenarios. In thiswork, we develop a hybrid (graph) neural network that combines a traditionalgraph convolutional network with an embedded differentiable fluid dynamicssimulator inside the network itself. By combining an actual CFD simulator (runon a much coarser resolution representation of the problem) with the graphnetwork, we show that we can both generalize well to new situations and benefitfrom the substantial speedup of neural network CFD predictions, while alsosubstantially outperforming the coarse CFD simulation alone.

 

Quick Read (beta)

loading the full paper ...