Home > TERATEC FORUM > Workshops > Workshop 4

Workshop 04 - 3:00 pm to 6:30 pm

Digital twins: the revolution is underway
Chaired by Pierre Louat, Senior Principal Application Engineer, Ansys and Cyril Baudry, Expert Senior HPC & Architecte SI Scientifique, EDF

Overview of digital twins usage at Michelin, applied to support the design of rubber manufacturing processes
By Thibault Dairay, Scientific Computing Expert, Michelin
and Vivien Clauzon, Architecte Solveur Elements Finis, Michelin

At the beginning of this presentation, we will explore a wide range of use cases for digital twins at Michelin. These applications will range from connected tires exchanging data with the vehicle to reduced order models designed to simulate tire behavior. These models can be provided to automotive manufacturers for virtual homologation.

Next, we will dive into a more detailed view of the simulation of rubber manufacturing processes. In particular, we will discuss how to combine data obtained from simulations with available measurements to create digital twins that accelerate the industrialization of our products.

Biography: Vivien is passionate about writing high-performance, maintainable scientific codes in answer to modern simulation needs. If it is clearly a technological and technical challenge, it is also an incredibly exciting and rich human adventure. He started his HPC journey back in 2004 while developing CFD codes able to scale on vectorial supercomputers during his PhD thesis. He then switched to GPGPU for some years and developed an interactive compressible, and later on incompressible, GPU-accelerated fluid flow solvers while conducting post-dotoral research. For 5 years at Numtech, he contributed to the development of resilient piloting systems for meteorology and air quality simulations, handled the internal HPC resources and took part in various academical partnerships. Working at Michelin since 2013 he dedicates most of his energy to the team in charge of developing the homemade Finite Elements solver and to a broader extent the service in charge of most of the simulation softwares in order to provide the most usefull, efficient and user friendly simulation tools.

Biography: Thibault is deeply passionate about navigating through diverse scientific and technical domains alongside R&D collaborators. His journey spans from understanding physical principles to their mathematical representations, and from traditional numerical techniques to cutting-edge machine learning algorithms. He firmly believes that true innovation arises from the synergy of multidisciplinary teams, comprising individuals with varied expertise, ranging from highly specialized professionals to practical end-users. His immersion into scientific computing began in 2010, during his doctoral research, where he implemented advanced numerical methods for simulating turbulent flows. Subsequently, he spent four years in academia, contributing to projects at institutions such as Imperial College London and Ecole Centrale de Lyon, focusing on modeling and simulation of turbulent flows. At Michelin since 2017, Thibault is working on subjects linked to several aspects of applied mathematics and numerical methods (modelling, optimization and control). His primary focus lies in the development of hybrid methodologies that bridge the gap between physics-based models and data-driven approaches. He firmly advocates for leveraging external knowledge sources, including the vibrant open-source community, recognizing their potential to enhance Michelin's simulation tools and methodologies. In his role, he leads academic and industrial partnerships to address the forthcoming challenges in Michelin's simulation initiatives.

Register now and get your badge here >>>