Search

Conférence: Confiance.ai Days, 4 au 6 octobre 2022, CentraleSupélec, Gif-sur-Yvette (France); This booklet gathers 52 papers, either in the form of articles or posters, presented during the secondedition of the Confiance.ai Days held in Saclay on October 4-6, 2022. Altogether they give a good snapshotof the research and development work done in the Confiance.ai program, an industrial and academicinitiative of the national Grand Challenge on provable and certifiable AI, launched in support of the France2030 strategy.Among these papers, a dozen are presented as « external contributions » that were selected by an ad-hoccommittee following a call for papers. All other communications belonged to one of five so-called « villages» with physical implementation in the conference hall, distributing the work done in Confiance.ai intofive topics : « End-to-end approach » ; « from Operational Design Domain to Data » ; « Explainabilityand Understanding» ; « Robustness and Monitoring » ; « Embedded AI ».After two years of activity, and complementing the Confiance.ai white paper, this document shows thediversity and the quality of the work done in the programme. Some important and up-to-date subjects areaddressed, such as – only to name a few - out-of-distribution detection, adversarial robustness, semi- orself-supervised learning, explainability by design, verification and validation, embedded AI etc. We hopethat you will enjoy reading parts of this document as much as we enjoyed preparing and attending the 2022Confiance AI days.

Topology optimization is devoted to the optimal design of structures. It aims at finding the best materialdistribution inside a working domain in order to satisfy mechanical, geometrical or manufacturingspecifications. The need for lighter and efficient structural solutions has made topology optimizationa vigorous research field in both academic and industrial structural engineering communities. In thepresent setting the level set method is coupled with a remeshing routine which enables the reconstructionof a body-fitted mesh at each step of the underlying optimization process. Since thestructural interface is known explicitely at each step of the iterative procedure, the body-fitted approachsimplifies the evaluation of the mechanical quantities of interest. The body-fitted remeshing strategy is driven by global user-defined parameters such as the minimal/maximal mesh size or thegradation of the desired mesh. In this contribution we propose to couple the body-fitted approach withadaptative remeshing techniques involving an anisotropic recovery-based a posteriori error analysis. Thegradient-recovery Zienkiewicz-Zhu-type error estimator is applied to drive isotropicor anisotropic mesh adaptation at each step of the underlying optimization process. The proposed errorestimator is used from one hand to control the accuracy of the discretization of the structural interface,from the other hand to control the discretization error linked to the numerical evaluation of mechanicalcriteria and sensitivities in order to speed up the overall optimization routine. All the proposed numericalexamples are realized using PISCO, a Research and Development software platform for shape andtopology optimization in active development at IRT SystemX. We rely on a gradient-flow optimizationalgorithm designed to decrease both the value of the objective function and the violation of the constraints.Several

Metal additive manufacturing (MAM) has grown, in recent years, very strong interest in academic researches as well as industrial applications. Among MAM processes, wire and arc additive manufacturing (WAAM) became very popular through its advantages in manufacturing of medium and large-scale components.The present work is carried out in the framework of the WAS project which deals with WAAM process. The process relies on an automatized welding process in which a part is built by successively deposed metal bead.We propose a physically based bead topology model using the equilibrium between the hydrostatic pressure and the capillarity force, under two-dimensional hypothesis. This equilibrium can be described by the Young-Laplace equation. The proposed model can also estimate a bead topology which is deposed on a complex support such as an inclined or a curved one. To do so, the Young equation is used to balance the forces at tri-phase point. Moreover, a deposed melted metal volume is necessary for the bead topology model. By modelling a gas metal arc welding (GMAW) power source system, the volume can be estimated and be used as a physical parameter for the bead topology model. Combining the topology and the power source models, the coupling model allows to simulate the topology of a part made of deposed beads via WAAM. In addition to the modelling, experimental profiles of the beads are used to validate the model.