Séminaire - Fabien Casenave : Maths-4-Innov-Action : Scientific Machine Learning for Industrial Design

Yvon Maday Informatique et sciences numériques Collège de France Année 2025-2026 Fabien Casenave Head of "Physics-informed machine learning and numerical experiments" team, Safran Senior Expert Séminaire - Fabien Casenave : Maths-4-Innov-Action : Scientific Machine Learning for Industrial Design Résumé In an industrial group such as Safran, numerical simulation of physical phenomena plays a central role in the design processes of manufactured products. At Safran Tech, the corporate research center of the Safran Group, we develop technologies to improve these processes by constructing fast and reliable surrogate models for a wide range of physical systems. This presentation introduces several technologies developed in recent years. First, we present a physics-based reduced-order modeling approach for nonlinear structural mechanics, applied to lifetime prediction of high-pressure turbine blades. Second, we address the learning of physics simulations under non-parameterized geometrical variability using classical machine learning techniques combined with nonlinear deterministic dimensionality reduction, including morphing, principal component analysis, and Gaussian process regression. We then illustrate the generation of mechanical components constrained by performance requirements. Finally, we highlight our contributions to the open-source and open-data Scientific Machine Learning community. Fabien Casenave Fabien Casenave est Senior Expert au département Digital Sciences and Technologies de Safran Tech, le centre de recherche du groupe Safran, où il dirige une équipe dédiée au Scientific Machine Learning. Ses travaux portent sur la réduction de modèles, les approches non intrusives et hybrides combinant apprentissage automatique et simulation physique, ainsi que sur le développement de bibliothèques open source pour la science et l'ingénierie numérique. Il a contribué à l'élaboration de méthodologies en réduction de modèles et en apprentissage pour la physique, avec des transferts vers des applications industrielles.