On the use of deep learning and computational fluid dynamics for the estimation of uniform momentum source components of propellers

• Published in: iScience 2023-12, Vol.26 (12), p.108297-108297, Article 108297
• Main Authors: Martínez-Cuenca, Raúl, Luis-Gómez, Jaume, Iserte, Sergio, Chiva, Sergio
• Format: Article
• Language: English
• Subjects: Artificial intelligence; Industrial engineering
• ISSN: 2589-0042; 2589-0042
• DOI: 10.1016/j.isci.2023.108297

This article proposes a novel method based on Deep Learning for the resolution of uniform momentum source terms in the Reynolds-Averaged Navier-Stokes equations. These source terms can represent several industrial devices (propellers, wind turbines, and so forth) in Computational Fluid Dynamics simulations. Current simulation methods require huge computational power, rely on strong assumptions or need additional information about the device that is being simulated. In this first approach to the new method, a Deep Learning system is trained with hundreds of Computational Fluid Dynamics simulations with uniform momemtum sources so that it can compute the one representing a given propeller from a reduced set of flow velocity measurements near it. Results show an overall relative error below the 5% for momentum sources for uniform sources and a moderate error when describing real propellers. This work will allow to simulate more accurately industrial devices with less computational cost. [Display omitted] •Propellers modeling as momentum sources is essential for industrial tanks design•A novel method to estimate source terms from velocity measurements was proposed•The method combines the power of Computational Fluid Dynamics and Neural Networks•Errors below 5% were found when inferring homogeneous momentum sources Artificial intelligence; Industrial engineering