ECCOMAS 2024

HORSES3D: a Machine Learning Accelerated High Order Discontinuous Galerkin Solver for CFD Simulations

  • Rubio, Gonzalo (UPM)
  • Ntoukas, Gerasimos (UPM)
  • Laskowski, Wojteck (UPM)
  • Mariño, Oscar (UPM)
  • Colombo, Stefano (UPM)
  • Mateo-Gabin, Andres (UPM)
  • Marbona, Himpu (UPM)
  • Otmani, Kheir-Eddine (UPM)
  • Manrique de Lara, Fernando (UPM)
  • Huergo, David (UPM)
  • Manzanero, Juan (Airbus)
  • Rueda-Ramirez, Andrés (University of Cologne)
  • Kopriva, David (FSU)
  • Valero, Eusebio (UPM)
  • Ferrer, Esteban (UPM)

Please login to view abstract download link

We present the latest developments of our open-source framework, HORSES3D (High-Order Spectral Element Solver). HORSES3D is a high-order discontinuous Galerkin solver designed to handle various flow applications, including compressible flows (with or without shocks), incompressible flows, different turbulence models (RANS and LES), particle dynamics, multiphase flows, and aeroacoustics. Recent updates allow us to efficiently simulate complex multiphysics scenarios, including turbulent flows, multiphase interactions, and moving bodies, through the utilization of local p-adaption and rapid multigrid time advancement. Additionally, we present recent research that integrates Machine Learning techniques with high-order simulations.