Robust Deep Learning For Emulating Turbulent Viscosities - Mise en forme des matériaux (CEMEF) Accéder directement au contenu
Article Dans Une Revue Physics of Fluids Année : 2021

Robust Deep Learning For Emulating Turbulent Viscosities

Aakash Patil
Aurélien Larcher
George El Haber
Elie Hachem

Résumé

From the simplest models to complex deep neural networks, modeling turbulence with machine learning techniques still offers multiple challenges. In this context, the present contribution proposes a robust strategy using patch-based training to learn turbulent viscosity from flow velocities, and demonstrates its efficient use on the Spalart-Allmaras turbulence model. Training datasets are generated for flow past twodimensional obstacles at high Reynolds numbers and used to train an auto-encoder type convolutional neural network with local patch inputs. Compared to a standard training technique, patch-based learning not only yields increased accuracy but also reduces the computational cost required for training.
Fichier principal
Vignette du fichier
POF21-AR-03388.pdf (4.77 Mo) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-03432657 , version 1 (17-11-2021)
hal-03432657 , version 2 (17-01-2023)

Identifiants

Citer

Aakash Patil, Jonathan Viquerat, Aurélien Larcher, George El Haber, Elie Hachem. Robust Deep Learning For Emulating Turbulent Viscosities. Physics of Fluids, 2021, 33 (10), pp.105118. ⟨10.1063/5.0064458⟩. ⟨hal-03432657v2⟩
130 Consultations
94 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More