Elhemmali, Alaaeddin, Huque, Mohammad Mojammel, Imtiaz, Syed, Azizur-Rahman, Mohammad et Ahmed, Salim.
2025.
« Experimental validation of pressure drop in turbulent annular flow using CFD and physical informed neural networks ».
In Proceedings of the CSME-CFDSC-CSR 2025 International Congress (Montreal, QC, Canada, May 25-28, 2025)
Coll. « Progress in Canadian Mechanical Engineering », vol. 8.
Prévisualisation |
PDF
37 - Experimental validation of pressure.pdf - Version publiée Licence d'utilisation : Tous les droits réservés aux détenteurs du droit d'auteur. Télécharger (1MB) | Prévisualisation |
Résumé
This work presents a comprehensive study in machine learning for fluid dynamics, with a focus on turbulent flow analysis and the application of Physics Informed Neural Networks (PINNs). Motivated by the need to validate PINNs against experimental data. This research investigates a lab scale setup that provides insights which can be extended to practical applications in the future. The experimental setup mimics annular flow in oil drilling were used for collecting flow and pressure data for validation then compared to the results of computational simulations that conducted using the Reynolds-Averaged Navier-Stokes (RANS) equations with the (k?? SST) turbulence model in ANSYS Fluent. Finally, PINNs were employed to solve governing equations directly, integrating physical laws and data without requiring turbulence models. The findings underscore the potential of PINNs as a resilient tool for fluid flow prediction, offering advantages in computational efficiency and simplicity over conventional CFD models.
| Type de document: | Compte rendu de conférence |
|---|---|
| Éditeurs: | Éditeurs ORCID Hof, Lucas A. NON SPÉCIFIÉ Di Labbio, Giuseppe NON SPÉCIFIÉ Tahan, Antoine NON SPÉCIFIÉ Sanjosé, Marlène NON SPÉCIFIÉ Lalonde, Sébastien NON SPÉCIFIÉ Demarquette, Nicole R. NON SPÉCIFIÉ |
| Date de dépôt: | 18 déc. 2025 15:10 |
| Dernière modification: | 18 déc. 2025 15:10 |
| URI: | https://espace2.etsmtl.ca/id/eprint/32392 |
Actions (Authentification requise)
![]() |
Dernière vérification avant le dépôt |

