CFC2023

MS1-RT Round Table - Data-driven methods and machine learning for CFD

Panelists: Laura Mainini (moderator), Elias Cueto, Matteo Diez, Luca Magri, Gianluigi Rozza, Tan Bui-Thanh, Andrea Beck, Francisco Chinesta
Wed, 26/04/2023
19:00 - 20:30
Rédaction 2
MS1-RT
Chaired by:
Dr. Laura Mainini (Imperial College London)

Panelists: Laura Mainini (moderator), Elias Cueto, Matteo Diez, Luca Magri, Gianluigi Rozza, Tan Bui-Thanh, Andrea Beck, Francisco Chinesta. Panel Goal: This Special Round Table aims at setting up a forum for the discussion of opportunities and open challenges introduced/offered by the use of data-driven methods (DDM) and machine learning (ML) for and with CFD problems. The forum brings together experts from applied mathematics, computational science, and engineering backgrounds to share their perspectives for a cross-fertilization of ideas and awareness of the state of the art. The main topics can be shaped as evolving along two orthogonal axes: (1) applications scaling from the use of DDM and ML to accelerate the solution of CFD problems to the use of DDM and ML to accelerate many-query problems –such as UQ and optimization for design, state/parameter identification, control– that are informed by the evaluation of CFD models; (2) validation and verification approaches to determine trustworthiness of the predictions ranging from a posteriori approaches to a priori or intrinsic characterization of the predictions in terms of interpretability, explainability, reliability and robustness. In addition to the discussion of the technical and scientific aspects, the need to broaden and improve people’s expertise in appropriately handling, and developing DDM/ML driven methods will be addressed. The diverse composition of the panel will be essential to draw the aspects common to different engineering applications as well as the most promising avenues explored by the community.