Fesquet, Jean, Alguacil, Antonio, Bousquet, Yannick, Sanjosé, Marlène, Moreau, Stéphane, Bauerheim, Michaël et Binder, Nicolas.
2026.
« Prediction of trailing edge wall-pressure fluctuations in rotor configurations with neural operators ».
In Proceedings of Global Power and Propulsion Society (GPPS Montreal26) (Montreal, QC, Canada, Sept. 09-11, 2026)
Prévisualisation |
PDF
Sanjose-M-2026-34614.pdf - Version publiée Licence d'utilisation : Creative Commons CC BY-NC-ND. Télécharger (8MB) | Prévisualisation |
Résumé
Rapid prediction methods to evaluate rotor self-noise can rely on turbulent boundary layer parameters extracted from Reynolds-Averaged Navier-Stokes (RANS) simulations, from which wall-pressure fluctuation spectra (WPS), the main noise sources, are reconstructed with models. An acoustic analogy then provides the far-field noise. To yield an efficient pipeline inside an aero-acoustic optimization framework over a large range of geometries, a recent state-of-the-art neural operator model (MARIO) that was shown to learn realistic turbomachinery flow fields is coupled with recently developed neural networks based either on Gene-Expressing Programming or an Autoencoder chained with a fully-connected Neural Network. MARIO has been retrained on RANS simulations of a mixed flow fan configuration at several operating conditions. MARIO is shown to accurately predict boundary layer parameters for wall-pressure spectra predictions. The newly developed MARIO+WPS pipeline has been successfully validated with accurate wall-pressure spectra extracted from a Large Eddy Simulation (LES) of the same configuration. Predictions of far-field noise based on Amiet’s approach and the newly developed data-driven WPS have also been shown to provide reliable estimates when compared with far-field noise obtained from LES using the Ffowcs Williams and Hawkings acoustic analogy.
| Type de document: | Compte rendu de conférence |
|---|---|
| Chercheur(-euse): | Chercheur(-euse) Sanjosé, Marlène |
| Affiliation: | Génie aérospatial |
| Date de dépôt: | 14 oct. 2026 18:47 |
| Dernière modification: | 16 oct. 2026 13:36 |
| URI: | https://espace2.etsmtl.ca/id/eprint/34614 |
Actions (Authentification requise)
 |
Dernière vérification avant le dépôt |