OATAO - Open Archive Toulouse Archive Ouverte Open Access Week

Towards a hybrid computational strategy based on Deep Learning for incompressible flows

Ajuria-Illaramendi, Ekhi and Alguacil, Antonio and Bauerheim, Michaël and Misdariis, Antony and Cuenot, Bénédicte and Benazera, Emmanuel Towards a hybrid computational strategy based on Deep Learning for incompressible flows. (2020) In: AIAA AVIATION FORUM, 15 June 2020 - 19 June 2020 (Virtual Event, United States).

[img]
Preview
(Document in English)

PDF (Author's version) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
3MB

Official URL: https://doi.org/10.2514/6.2020-3058

Abstract

The Poisson equation is present in very different domains of physics and engineering. In most cases, this equation can not be solved directly and iterative solvers are used. For many solvers, this step is computationally intensive. In this study, an alternative resolution method based on neural networks is evaluated for incompressible flows. A fluid solver coupled with a Convolutional Neural Network is developed and trained on random cases with constant density to predict the pressure field. Its performance is tested in a plume configuration, with different buoyancy forces, parametrized by the Richardson number. The neural network is compared to a traditional Jacobi solver. The performance improvement is considerable, although the accuracy of the network is found to depend on the flow operating point: low errors are obtained at low Richardson numbers, whereas the fluid solver becomes unstable with large errors for large Richardson number. Finally, a hybrid strategy is proposed in order to benefit from the calculation acceleration while ensuring a user-defined accuracy level. In particular, this hybrid CFD-NN strategy, by maintaining the desired accuracy whatever the flow condition, makes the code stable and reliable even at large Richardson numbers for which the network was not trained for. This study demonstrates the capability of the hybrid approach to tackle new flow physics, unseen during the network training.

Item Type:Conference or Workshop Item (Paper)
HAL Id:hal-02923501
Audience (conference):International conference proceedings
Uncontrolled Keywords:
Institution:Université de Toulouse > Institut Supérieur de l'Aéronautique et de l'Espace - ISAE-SUPAERO (FRANCE)
Other partners > Centre Européen de Recherche et Formation Avancées en Calcul Scientifique - CERFACS (FRANCE)
Other partners > Jolibrain (FRANCE)
Laboratory name:
Statistics:download
Deposited On:27 Aug 2020 08:33

Repository Staff Only: item control page