← Papers

ML-accelerated CFD

Machine learning accelerated computational fluid dynamics

Paper title Machine learning accelerated computational fluid dynamics

Kochkov et al. (Google Research) · Aero & CFD

Paper arXiv 2102.01010PDF

Keeps a finite-volume Navier–Stokes solver and replaces only the part most hurt by a coarse grid: a convolutional network, trained end to end through the differentiable solver written in JAX, learns how to interpolate the convective fluxes on the coarse mesh.

Published
PNAS

Share on LinkedIn, X or Bluesky, or or .

Know related work that belongs here?

Something wrong on this page?

More papers

All papers
  1. PhysicsNeMo-Mesh — GPU-Native Meshes and Simulation Cases for Physics AI
  2. Multi-fidelity scaling — Compute-optimal data scaling via multi-fidelity training
  3. CAE-UQ — Predictive Uncertainty for Neural CAE Surrogates