Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 3
Encontrada(s) 1 página(s)

Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Active flow control of a turbulent separation bubble through deep reinforcement learning

  • Font, Bernat
  • Alcántara Ávila, Francisco
  • Rabault, Jean
  • Vinuesa, Ricardo
  • Lehmkuhl Barba, Oriol
Journal of Physics: Conference Series 5th Madrid Turbulence Workshop 29/05/2023 - 30/06/2023 Madrid, Spain, The control efficacy of classical periodic forcing and deep reinforcement learning (DRL) is assessed for a turbulent separation bubble (TSB) at Reτ = 180 on the upstream region before separation occurs. The TSB can resemble a separation phenomenon naturally arising in wings, and a successful reduction of the TSB can have practical implications in the reduction of the aviation carbon footprint. We find that the classical zero-net-mas-flux (ZNMF) periodic control is able to reduce the TSB by 15.7%. On the other hand, the DRL-based control achieves 25.3% reduction and provides a smoother control strategy while also being ZNMF. To the best of our knowledge, the current test case is the highest Reynolds-number flow that has been successfully controlled using DRL to this date. In future work, these results will be scaled to well-resolved large-eddy simulation grids. Furthermore, we provide details of our open-source CFD–DRL framework suited for the next generation of exascale computing machines., This work was performer in part during the Fifth Madrid Summer Workshop, funded by the European Research Council (ERC) under the Caust grant ERC-AdG-101018287. OL has been partially funded by the European Commission’s Horizon 2020 Framework program and the European High-Performance Computing Joint Undertaking (JU) under grant agreement no. 101093393, and by MCIN/AEI/10.13039/501100011033 and the European Union NextGenerationEU/PRTR (PCI2022-134996-2), project CEEC. OL has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I). RV and FAA acknowledge financial support from ERC grant no. ‘2021-CoG-101043998, DEEPCONTROL’. The authors thank Laia Juli´o from the BSC support services and Dr. Andrew Shao from the SmartSim team for the help and fruitful discussions about the CFD–DRL software framework. The authors also thank the UPM Turbulence Summer School and our hosts Dr. Miguel P´ erez Encinar and Dr. Adri´an Lozano-Dur´an for the organization of the workshop. Finally, we acknowledge the National Academic Infrastructure for Supercomputing in Sweden (NAISS) for the computational time provided at the large-GPU systems Alvis and Berzelius., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Towards a machine learning model for explicit algebraic Reynolds stress modelling using multi-expression programming

  • Miró Jané, Arnau
  • Wallin, Stefan
  • Colombo, Alessandro
  • Temmerman, Lionel
  • Wunsch, Dirk
  • Lehmkuhl Barba, Oriol
A novel framework for a 5-term explicit algebraic Reynolds stress model has been proposed, which is suitable for machine learning procedures. Under this framework, high-fidelity datasets have been explored and post-processed. Through multi-expression programming, models have been obtained using these data. They have been thoroughly tested and present all the desirable features of a model. Implementation of these models in a numerical code has also been possible. Results have shown that they present a small upgrade over classical models for certain kinds of flows, thus showing that the physics have been captured well. It is proved that the present workflow can obtain valid expressions with similar to better performance than the baseline model., The research leading to this work has received funding within the HiFi-TURB project from the European Union’s Horizon2020 research and innovation programme under grant agreement No 814837. Oriol Lehmkuhl has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Towards the usage of geometry agnostic ß-variational-autoencoders for model order reduction of turbulent flows

  • Eiximeno Franch, Benet
  • Miró Jané, Arnau
  • Kumar, Vishal
  • Rodríguez Pérez, Ivette María
  • Lehmkuhl Barba, Oriol
The present work introduces a novel data compression framework for model order reduction of turbulent flows. The compression algorithm relies on the Geometry Agnostic Variatonal-autoencoders Integration (GAVI) technique. GAVI decouples the geometry and mesh information from the process of reducing a dataset via a CNN-based autoencoder via the QR factorization of the snapshot matrix. The framework is tested on a channel flow at Ret = 950. After reducing 9 times the amount of storage required, the root mean square of the velocity fluctuations and the budgets of the Reynolds stresses are recovered with excellent agreement when compared to the original DNS., The research leading to this work has been partially funded by the project TIFON with reference PLEC2023-010251/ AEI/10.13039/501100011033. B. Eiximeno’s work was funded by a contract from the Subprograma de Ayudas Predoctorales given by the Ministerio de Ciencia e Innovación (PRE2021- 096927). O. Lehmkuhl has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I). V. Kumar acknowledges his AI4S fellowship within the Generación D initiative by Red.es, Ministerio para la Transformación Digital y de la Función Pública, for talent attraction (C005/24-ED CV1), funded by NextGenerationEU through PRTR. The authors acknowledge the support of the Departament de Recerca i Universitats de la Generalitat de Catalunya through the research group Large-scale Computational Fluid Dynamics (ref.: 2021 SGR 00902) and the Turbulence and Aerodynamics Research Group (ref.: 2021 SGR 01051)., Peer Reviewed, 9 - Indústria, Innovació i Infraestructura, Postprint (published version)