Publications

Physics

Data-driven Fluid Flow Prediction Using Conditional Score-based Diffusion Models

Advances in computational power have made turbulent field simulations central to many disciplines, yet traditional solvers struggle to deliver fast flow estimations due to the inherent chaotic nature of the problem. More recently, diffusion models…

Wilfried Genuist, Éric Savin, Filippo Gatti, Didier Clouteau

AICOMAS 2025 - Artificial Intelligence and computational Methodes in Applied sciences - 2025

Matériaux composites et construction

Carbon nanotube effects on low-cement mortar performance and eco-efficiency

The construction sector, particularly cement production, accounts for 8-10 % of global CO2 emissions annually, and the industry is under increasing pressure to adopt sustainable practices, with carbon nanotubes (CNTs) emerging as a promising…

E.D. Reis, Fabrice Gatuingt, F.S.J. Poggiali, A.C.S. Bezerra

Journal of Building Engineering - 2025

Structural mechanics

Approches local/global non-intrusives et optimisation robuste

Les approches de submodeling sont très prisées par les industriels car elles permettent de réanalyser à faible coût, au sein de logiciels commerciaux, des zones locales jugées critiques en raffinant les maillages et la description de la géométrie ou…

Léa Karaouni, Ludovic Chamoin, Olivier Allix, S. Guinard

26ème Congrès Français de Mécanique 2025 - 2025

Materials and structures in mechanics

Unified kinetic continuous damage model for random LCF-HCF interaction

Amélie Fau, Mainak Bhattacharyya, Rodrigue Desmorat

European Journal of Mechanics - A/Solids - 2025

Machine Learning

Autoregressive regularized score-based diffusion models for multi-scenarios fluid flow prediction

Building on recent advances in scientific machine learning and generative modeling for computational fluid dynamics, we propose a conditional score-based diffusion model designed for multi-scenarios fluid flow prediction. Our model integrates an…

Wilfried Genuist, Éric Savin, Filippo Gatti, Didier Clouteau

Nonlinear Sciences

Extension of quasi-static damage meso-modelling of composite laminates to dynamic failure and erosion

Impact tests on structures are destructive, costly and difficult to analyze. This is particularly true for composite structures, whose properties and failure mechanisms are highly dependent on the chosen stacking architecture. As a result, virtual…

O. Allix

Plenary Lecture, CFRAC 2025 - The 8th International Conference on Computational Modeling of Fracture and Failure of Materials and Structures - 2025

Artificial Intelligence

Accelerating nonlinear finite element analysis via residual-aware neural network constitutive models

Nonlinear finite element analysis (FEA) relies heavily on iterative methods such as the Newton-Raphson algorithm, with computational cost primarily driven by the repeated solution of large linear systems (global stage) and the evaluation of…

Pierre-Eliot Malleval, Victor Matray, Faisal Amlani, Ronan Scanff, Frédéric Feyel, David Néron

Finite Elements in Analysis and Design - 2025

Physics

Score-Based Diffusion Models with Autocorrelation Regularization for Fluid Flow Prediction

The simulation of turbulent flows is an essential field of study with widespread applications. While traditional solvers struggle to provide rapid estimations of solutions due to the intrinsic chaotic nature of turbulence, recent breakthroughs in…

Wilfried Genuist, E Savin, Filippo Gatti, D Clouteau

2025 IACM Computational Fluids Conference (CFC) - 2025

Engineering Sciences

Recueil de communications école d'hiver 2025 GDR ARCHI-META - DN METACMED

The last 20 years have seen a remarkable growth in scientific interest in (elastic and mechanical) metamaterials. One of the reasons for this is the great potential that this subject area has in itself as a “link between different disciplines”,…

Marco Miniaci, Giuseppe Rosi, Nicolas Auffray, Cédric Bellis, Christelle Combescure, Justin Dirrenberger, Kerem Ege, Muamer Kadic, Fabrice Lemoult, Bruno Morvan, Vincent Pagneux, Julien Poittevin, Martin Poncelet

Ecole d'hiver GDR ARCHI META - DN METACMED - 2025

Engineering Sciences

Comparative Hygrothermal Performance of a Mortar Wall with Micronized Miscanthus Fibers versus a Reference Wall: Experimental and Numerical Analysis

Miscanthus fibres are regarded as a sustainable, eco-friendly material with high hygroscopicity and low thermal conductivity. Although promising at the material scale, further wall-scale evaluations are needed. This study investigates the…

Haoyan Zhang, Franck Komi Gbekou, Karim Benzarti, Rachid Bennacer, Abderrahim Boudenne

Journal of Building Engineering - 2025