Marcello Meldi a obtenu son Master en aérospatial à l'Université de Pise en 2007 et a poursuivi ses études à l'Université de Modena et à Reggio Emilia, où il a obtenu son doctorat en conception et technologie automobile en 2011. Il a ensuite déménagé en France où il a travaillé comme postdoctorant dans deux laboratoires différents (Institut d'Alembert, Paris et M2P2, Marseille) avant d'être recruté comme maître de conférences à l'ISAE-ENSMA, Poitiers en 2014. Il y effectue son enseignement au département d'aérodynamique et de mathématiques et développe ses recherches au sein de l'Institut Pprime. Très récemment, il a été recruté comme professeur à l'ENSAM - Campus Lille, où il prendra ses fonctions à partir de janvier 2022. L'intérêt de Marcello Meldi pour la recherche porte sur la simulation numérique des écoulements turbulents. Ses travaux dans ce domaine visent la représentation réaliste de la turbulence en combinant des techniques numériques avec des outils séquentiels pilotés par les données, tels que le filtre de Kalman.
Marcello Meldi received his Master's degree in Aerospace Engineering at University Pisa in 2007 and continued his studies at University of Modena and Reggio Emilia, where he obtained his PhD in Automotive Design & Technology in 2011. He then moved to France where he worked as a postodoc in two different laboratories (Institut d'Alembert, Paris and M2P2, Marseille) before being recruited as assistant professor at ISAE-ENSMA, Poitiers in 2014. There, he performed his teaching tasks in the department of aerodynamics and mathematics and he developed his research within the Pprime Institute. Very recently, he has been recruited as full professor at ENSAM - Campus Lille, where he will start his duties from January 2022. Marcello Meldi's interest in research deals with the numerical simulation of turbulent flows. His work in this field aims for the realistic representation of turbulence combining numerical techniques with data-driven sequential tools, such as the Kalman Filter.
Sequential Data Assimilation methods for the analysis of unsteady / turbulent flows
Abstract: The constant increase in computer power and new emerging computational architectures are providing unprecedented opportunities for the numerical analysis of complex flow configurations. Within this framework, the representation of turbulent flows and their sensitivity to realistic epistemic uncertainties affecting their evolution is one of the most difficult open challenges. Recently, the scientific community has integrated data driven techniques to improve the prediction of such flow configurations. Will these tools be able to provide a breakthrough in the way we analyse turbulent flows? In the present talk, an overview of recent works developed by the research team on the field of Data Assimilation (DA) is proposed. DA is a science usually associated with Bayesian estimation which is devoted to optimal flow reconstruction, in a probabilistic sense, starting from different sources of information which are characterized by a level of confidence. The presentation of the results obtained by the team will be enriched by discussion about the achievements obtained with these tools in the state of the art as well as indicating future perspectives of application for the analysis of complex configurations
25 November 2021, 16h3017h30
Veuillez contacter F. Romano pour obtenir le lien
