Learning to Fly: Harnessing High-Performance Computing and Machine Learning for Sustainable Aviation

Advances in high-performance computing (HPC) enable the simulation of complex, multiscale flow phenomena relevant to aerospace with unprecedented accuracy, generating vast high-fidelity datasets. In parallel, scientific machine learning (SML) is rapidly transforming the physical sciences. When combined, SML and HPC promise a step change in predictive capabilities and computational efficiency for CFD, with potential impact on decarbonizing aviation and other carbon-intensive sectors. Yet, key challenges remain: selecting informative training data, generalizing to out-of-distribution conditions, and quantifying predictive uncertainty. In this talk, I will explore how the synergy between HPC and SML can be harnessed to accelerate sustainable aviation, and highlight current bottlenecks and research directions toward trustworthy, scalable ML-enhanced CFD.

Plenary Speakers

Paola Cinnella (Sorbonne University, France)

Paola Cinnella is Professor of Fluid Mechanics at the Jean Le Rond d’Alembert Institute, Sorbonne University. Her research focuses on Computational Fluid Dynamics, including discretization methods, optimization, uncertainty quantification, and scientific machine learning for turbulent flows in aerospace and energy. She is Editor-in-Chief of Computers & Fluids, Associate Editor of International Journal of Heat, and serves on several editorial boards. She also coordinates the “Machine Learning in Fluid Dynamics” group of ERCOFTAC, the European research community in Flow, Turbulence, and Combustion.

Matej Praprotnik (National Institute of Chemistry, Slovenia)

Matej Praprotnik is Head of the Laboratory for Molecular Modeling at the National Institute of Chemistry and Professor of Physics at the Faculty of Mathematics and Physics, University of Ljubljana. He is a former Chair of the PRACE Scientific Steering Committee and serves as a council member of CECAM. Matej is a recipient of the ERC Advanced Grant 2019 by the European Research Council and the project coordinator of MultiXscale, a EuroHPC JU CoE. His research is focused on computer simulation of soft and biological matter. The focus is on developing and combining innovative computational and theoretical methods augmented by machine learning techniques to study complex molecular systems.