Ludovic Chamoin (Professeur, Ecole Normale Supérieure Paris-Saclay) a animé un Seminar@SystemX en webinaire sur le thème « Numerical strategies mixing models and data for the real-time monitoring of complex mechanical systems », le 17 juin 2021 de 14h à 15h30.
Résumé (en anglais) :
The presentation aims at introducing effective numerical tools that permit real-time and robust data assimilation and control on mechanical structures. It is placed in the framework of the DDDAS (Dynamic Data Driven Application System) technology in which a continuous exchange between simulations and experimental measurements is implemented in order to create a retroactive control loop and design smart structures. In this context, and in order to take various uncertainty sources (modeling error, measurement noise,..) into account, a general stochastic methodology with Bayesian inference may be considered. It then needs to be coupled with advanced numerical techniques (model reduction, Transport Map sampling) in order to decrease the computational complexity and make the strategy compatible with the real-time constraint. Alternative robust strategies may also be envisioned, such as using Kalman filtering together with deterministic approaches that incorporate physical information on the specific structural behavior. In any case, addressing efficiency and reliability with complex mechanical systems also requires to resort to adaptive modeling (to compute right at the right cost, with smart and dynamical management of computing resources depending on the objective and experimental information), and to data-based model enrichment through physics-guided learning.
The investigated numerical methodologies will be illustrated on several practical applications, such as the online control of welding processes or the online health monitoring a damageable structures. Also, new challenges and opportunities will be discussed in association with a starting ERC project on the topic.
Biographie (en anglais) :
Ludovic Chamoin is a Full Professor at Ecole Normale Supérieure Paris-Saclay. He received his PhD from ENS Cachan in 2007, under the supervision of P. Ladevèze. After a postdoctoral stay in the J.T. Oden’s group at the University of Texas at Austin, he came back in France as an Assistant Professor. He then took a two-year leave at INRIA Paris before joining his current position in 2016.
The research activities of Ludovic Chamoin span various fields of computational mechanics, with a strong link with applied mathematics and engineering. Topics include model certification and adaptation, inverse problems, model validation, multiscale analysis, model reduction, uncertainty quantification, and optimal control applied to structural mechanics. Over the last years, he has been the Head of a research team working on real-time data assimilation and system control, with several research collaborations with industry and international teams.
Ludovic Chamoin published about 80 papers in international journals and received several awards, including the John Argyris award, the Robert J. Melosh medal, or the CSMA PhD award. He was an invited lecturer at Collège de France (2014), at the ADMOS conference (2017), and at the World Congress in Computational Mechanics (2020). He served as President of the CSMA Juniors section (young researchers of the CSMA French association) that he created in 2017, and as a representative in the ECCOMAS Young Investigators Committee. In 2019, he was elected Junior Member of the Institut Universitaire de France (IUF). In 2020, he obtained an ERC-Consolidator grant from the European Research Council on a research project (starting in June 2021) dealing with real-time control of structural damage.
