HAL-MINES ParisTech
Not a member yet
27348 research outputs found
Sort by
High-fidelity modeling of interface crossing in the diffusion welding process at the polycrystalline scale
No full textControlling the microstructure of a diffusion welded interface is a critical point to ensure optimum mechanical properties and the homogeneity of the joint. Beyond the intimate contact formation between bonded parts studied in the literature, this article focuses on the grain boundary crossing of the interface during this process and its measurement. Following this perspective, a Level-Set method has been used for full-field microstructure simulations in 2D with various interface parameters. Two crossing measurement models have been formulated, tested and discussed over the simulations
AtomSurf : Surface Representation for Learning on Protein Structures
No full textInternational audienceWhile there has been significant progress in evaluating and comparing different representations for learning on protein data, the role of surface-based learning approaches remains not well-understood. In particular, there is a lack of direct and fair benchmark comparison between the best available surface-based learning methods against alternative representations such as graphs. Moreover, the few existing surface-based approaches either use surface information in isolation or, at best, perform global pooling between surface and graph-based architectures. In this work, we fill this gap by first adapting a state-of-the-art surface encoder for protein learning tasks. We then perform a direct and fair comparison of the resulting method against alternative approaches within the Atom3D benchmark, highlighting the limitations of pure surface-based learning. Finally, we propose an integrated approach, which allows learned feature sharing between graphs and surface representations on the level of nodes and vertices across all layers. We demonstrate that the resulting architecture achieves state-of-the-art results on all tasks in the Atom3D benchmark, while adhering to the strict benchmark protocol, as well as more broadly on binding site identification and binding pocket classification. Furthermore, we use coarsened surfaces and optimize our approach for efficiency, making our tool competitive in training and inference time with existing techniques. Code can be found online: https://github.com/Vincentx15/atomsur
Effet d'une pré-déformation plastique sur le comportement en fatigue d'un acier martensitique à haute résistance
No full textInternational audienceThis study investigates the impact of prior plastic strain deformation on the residual fatigue lifetime and crack mechanisms of a high-strength martensitic steel. Polished low-cycle fatigue specimens were tested with and without prior tensile deformation up to 4%. Additional specimens were cut from components that underwent cold forming and stress relieving. Manson-Coffin curves were determined, and damage mechanisms were analyzed using 2D and 3D techniques. All specimens exhibited cyclic softening; the evolutions of stress amplitudes with the current number of cycles merged into a single curve until damage development. Prior tensile deformation reduced fatigue lifetime by approximately 10%, without altering the Manson-Coffin exponent, and the cumulative plastic strain at fracture also decreased. These effects were consistent regardless of prior deformation conditions. In a given specimen, several fatigue cracks initiated from persistent slip bands and propagated independently, extending toward the specimen axis up to 25% of the specimen radius. Comparison between cold-formed and tensile prestrained specimens showed that prior tensile plastic deformation proved effective in assessing the residual low cycle fatigue strength of cold deformed highstrength martensitic steel
Microstructural evolution of AA7075 powder as feedstock material for cold spray deposition
No full textInternational audienc
DOC-Depth: A Novel Approach for Dense Depth Ground Truth Generation
No full textInternational audienc
Expression de l'Optimum Climatique de l'Éocène inférieur en domaine littoral et continental : exemple du Bassin de Paris
No full textInternational audienceL’Optimum Climatique de l’´Eocène inférieur ou Early Eocene Climatic Optimum (EECO) s’étend d'environ 53 à 49 Millions d’années (Ma). Il correspond à la période la plus chaude du Cénozoïque avec des températures de 10 à 15°C supérieures à l’Actuel et une pCO2 maximale d’environ 1600 ppm aux environs de 51 Ma. L’EECO a été très bien défini en domaine marin grâce à l’analyse géochimique de foraminifères benthiques et planctoniques. À l’inverse, cetanalogue du scénario SPP5-8.5 du GIEC reste beaucoup moins contraint en domaine littoral et continental. Le Bassin de Paris donne accès à des dépôts de ces environnements encore peu connus d’un point de vue paléoclimatique, contemporains de l’EECO et avec un contenu fossilifère abondant et très bien préservé. Ainsi, l’étude géochimique des fossiles de mollusques dans ces sédiments permet de contraindre l’expression de cet optimum sur la variabilité climatique locale des températures.Plusieurs sites répondant à ces critères ont été ciblés dans le Bassin de Paris. Des échantillons ont pu être prélevés sur les sites de Cuise-la-Motte (stratotype du Cuisien ; ancien Yprésien supérieur) qui présente des dépôts marins proximaux, de Mancy avec des dépôts fluviatiles et du Grand Séminaire de Soissons avec une transition de dépôts fluvio-lacustres à littoraux. La sédimentation sableuse à argilo-sableuse que l’on retrouve sur ces différents sites et leur faible enfouissement ont permis une préservation idéale des fossiles, confirmée par des observationsen cathodoluminescence des bivalves marins et d’eau douce.Des analyses isotopiques en δ18O et δ13C ont pu être réalisées sur plusieurs spécimens d’environnements variés tels que des huîtres (littoral) ou des unios (eau douce). Ces premières analyses fournissent des données pour le début de l’´Eocène (échantillons de Soissons) et permettent de comparer les paléotempératures en contexte littoral et continental pour la fin de l’Yprésien (échantillons de Cuise-la-Motte et Mancy). Ces résultats seront complétés parla suite par des analyses en Δ47 qui devraient également donner accès aux variations de salinité et mieux contraindre les paléoenvironnements et le régime hydrologique de l’époque
Multi-Level Feature Fusion for Continual Learning in Visual Quality Inspection
No full textInternational audienc
Modeling the influence of the vadose zone in a geophysical study of the ORGEVAL critical zone observatory
No full textInternational audienceQuantifying the water and heat fluxes at the interface between surface and groundwater (GW) includingthe vadose zone (VZ), surface water (SW) and GW interfaces is a key issue for water and energy resourcemanagement. This is because it ensures a safe yield and good water quality under the ongoing global change.Such quantification with field measurements is not straightforward because the SW-GW changes depend onthe boundary conditions and the spatial description of the hydrofacies. These are not well known and areusually estimated by calibrating models using classical data like hydraulic heads and river discharge. Toovercome the lack of direct in situ data, de Marsily et al., 2005 advice to couple the classical observation withunconventional data such as the geophysical surveys (Binley et al., 2015). This study provides new insightabout the use of the geophysical methods, such as heat as a tracer, electrical resistivity tomography (ERT)and seismic monitoring to characterize the VZ functionning. We develop a methodology to build strongerconstraints to the inversion methods to estimate the hydrodynamic and thermal parameter and the boundaryconditions, both in space and time, by using a multi-method approach. The hydrogeological forward modelGinette (Rivière et al., 2016), based on the Richards equation and the heat transport. This forward modelis used to study variability saturated flow and the heat fluxes in 1D heterogeneous soil subject to seasonalvariation in precipitation and temperature. The sensitivity of geophysical data on the VZ parameters isestimated by using synthetic cases and the SantiLudo model. The SantiLudo model (Solazzi et al., 2021) isfounded on the Hertz-Mindlin contact theory combined with the Biot-Gassmann model and the influence ofcapillary succion. This model is used to obtain the mechenical hence siesmic properties of a soil in relationto its degree of saturation. The synthetic case study developed, that mimics the VZ of the Orgeval criticalzone observatory (France) influenced by temperate climatic conditions and hydrological forcings, considersthe hydaulic and thermal parameters of various porous media representative of a wide range of hydrofacies
From theory of dendrite tip kinetics to additive manufacturing of multicomponent alloys
No full textInternational audienc
Computational Modeling and Trends in RAI Therapy for Thyroid Cancer: A Field with Clinical Promise
No full textInternational audienceRadioactive iodine (RAI) therapy remains a pillar in the treatment of differentiated thyroid cancer, particularly in metastatic or high-risk cases. However, inter-individual variability in therapeutic response and the risk of overtreatment poses enduring clinical challenges. In parallel, a growing body of research seeks to optimize RAI protocols through computational modeling, artificial intelligence (AI), and personalized simulation tools. To assess this emerging field, a natural language processing (NLP) pipeline that aggregates and analyzes bibliographic data (from Scopus, CrossRef, and OpenAlex) was developped. The system filters publications based on conceptual relevance using MiniLM embeddings and thematic keyword detection, enabling structured classification by methodology, geographic origin, and modeling framework