Publications

This study investigates the decomposition of retained austenite (RA) in tool steels for plastic molding in correlation with the alloy chemical composition and the tempering parameters. Two grades differing in their silicon content with initial mixed bainitic/martensitic microstructures were investigated using in situ synchrotron high-energy X-ray diffraction (HEXRD) during tempering in the 550 °C to 600 °C temperature range for one-hour holding time. Results indicated carbide formation during heating or isothermal holding; however, retained austenite remained untransformed up to the end of the tempering holding time in all investigated conditions for both grades. In situ HEXRD provides direct evidence of the transformation of retained austenite into fresh martensite on cooling from the tempering stage. This behavior is correlated to the evolution of carbon enrichment of retained austenite and the effect of silicon is discussed.

In machine-tools, geometrical defects are unavoidable. They can greatly affect the dimensional accuracy of the final workpiece if not corrected. Software compensation strategies are less expensive than mechanical adjustments and they provide great improvement in volumetric accuracy. In this study, different compensation methods are compared in a 3-axis milling applications: Numerical Controller (NC) internal compensation tables, modification of the programmed tool-path (G code) and modification of position feedback signals. The latter is the main purpose of this work, because it shows great potential and is not linked to one particular type of NC. It communicates with a custom software application that processes the position data and generates corrected signals according to a geometric model based on the rigid body assumption. The NC is then induced to perform volumetric error correction based on its default programming. The compensation methods are compared based on their ability to bring out or correct imposed geometric errors. The highlighted solution shows performances comparable to the G-code modification by correcting more than 96% of the imposed geometric errors without affecting the numerical chain from the program generation to its execution on the machine. It is also independent of the NC or the motors control cards.

This article presents the architecture of MemorIA, an integrative system that combines existing AI technologies into a coherent educational framework for creating interactive historical agents, with the aim of fostering students' learning interest. MemorIA generates animated digital portraits of historical figures, synchronizing facial expressions with synthesized speech to enable natural conversations with students. The system leverages NVIDIA Audio2Face for real‐time facial animation with first‐order motion model for portrait manipulation, achieving fluid interaction through low‐latency audio‐visual streaming. To assess our architecture in a field situation, we conducted a pilot study in middle school history classes, where students and teachers engaged in direct conversation with a virtual Julius Caesar during Roman history lessons. Students asked questions about ancient Rome, receiving contextually appropriate responses. While qualitative feedback suggests a positive trend toward increased student participation, some weaknesses and ethical considerations emerged. Based on this assessment, we discuss implementation challenges, suggest architectural improvements, and explore potential applications across various disciplines.

This work offers a detailed examination of the phase field approach for modeling brittle fracture, emphasizing its theoretical foundations, mathematical descriptions, and computational strategies. Central to our discussion is an in-depth analysis of strain energy decomposition methods integral to phase field models. We introduce an innovative technique using a cleavage plane based degradation that has shown promising results under various loading scenarios. We meticulously evaluate each method’s inherent limitations and challenges to highlight their respective advantages and drawbacks across different loading scenarios. This review aims not only to catalog existing knowledge but also to pave the way for future research directions in the application of phase field approach to fracture analysis.

A method of generating a machining program with a programming system comprising a programming module and a man-machine communication interface having a display surface and selection means. The method comprises : - displaying on the display surface an image illustrating a representation of a real part to be machined and a virtual model corresponding to the real part to be machined;- selecting in the image, using the selection means, at least one given virtual zone of the predetermined virtual model corresponding to a given real zone of the part to be machined and a desired machining intensity parameter for the selected virtual zone; then- generating the machining program comprising a machining intensity characteristic to be applied as a function of the selected parameter.

Procédé de génération d'un programme d'usinage avec un système de programmation comprenant un module de programmation et une interface de communication homme-machine comportant une surface d'affichage et des moyens de sélection. Le procédé comprend : - l'affichage sur la surface d'affichage d'une image illustrant une représentation d'une pièce réelle à usiner et d'un modèle virtuel correspondant à la pièce réelle à usiner ; - la sélection dans l'image, à l'aide des moyens de sélection, d'au moins une zone virtuelle donnée du modèle virtuel prédéterminé correspondant à une zone réelle donnée de la pièce à usiner et d'un paramètre d'intensité d'usinage souhaité pour la zone virtuelle sélectionnée ; puis - la génération du programme d'usinage comprenant une caractéristique d'intensité d'usinage à appliquer en fonction du paramètre sélectionné.

Dissipative components of tool-workpiece interaction are of major importance in cutting-related vibrations. At the macroscopic vibration scale, such dissipation is usually accounted for by additional generalized damping forces in the equation of motion of the system’s elastodynamics. A finer consideration at cutting edge scale would bring up a line-distributed force mostly of ploughing nature. These two scales are usually linked by analytical integration, involving simplifying kinematical assumptions. In the present work a comparative investigation is proposed, for a machining operation, considering both representations in a detailed time domain modeling framework. Tool’s cutting edges are represented in a discretized manner, i.e. split into numerous elementary cutters allowing for detailed tool-workpiece interaction force distribution. The matter removal process is modeled via dexel-based surface discretization coupled with finite element-based modal shapes, enabling a consistent machined surface generation representation. Finally, the equations of motion are formulated for modal degrees of freedom and solved by a time marching algorithm. Based on these analyses, the limitations of resulting process damping force terms representations are considered regarding vibrations and interaction force magnitudes.

The austenitic stainless steels are characterised by remarkable mechanical properties, combining high ductility and high strength. Their plastic deformation involves a wide variety of strain induced deformation mechanisms, strongly related to the alloy stacking fault energy. With increasing SFE, martensitic transformation sequences (gamma to alpha', gamma to epsilon to alpha'), deformation twinning, glide of dissociated or perfect dislocations can occur. A quantitative modelling of such a deformation behaviour, with real predictive capabilities, has been developed recently. It is formulated in terms of finite strains and takes the various inelastic strains encountered in the material into account. In particular, two inelastic deformations are considered, either epsilon martensite transformation strain or twinning strain, depending on the testing temperature, and the alpha' transformation strain. Thermomechanical couplings are realised between these two inelastic deformation modes. The model which uses a self-consistent method for the scale transition, allows us to calculate the global behaviour of the polycrystal. In this contribution, this new model is applied to foresee the behaviour of a 304 stainless steel, tensile tested in a temperature range of -60 degrees Celcius to room temperature. Besides, large experimental investigations were performed on various 304 specimens. In particular, X-ray diffraction was used to quantify the volume fraction of the alpha' and epsilon martensite and to determine the texture evolution of the parent and the product phases. The local microtextures were analysed with the EBSD technique in a SEM FEG. The predictions obtained with the micromechanical model are compared to the different experimental results, notably the mechanical behaviour and associated deformation mechanisms, the transformation kinetic as well as the texture evolution.

Geometric errors in amachine tool structure are mainly responsible for the volumetric error in theworkspace. They occur at the attachment of each link between axis joints, but also along each axis in their joint frame. Reducing the impact of these errors is a key factor in guaranteeing the functional requirements of high value-added parts. Unlike mechanical correction, software compensation strategies are often chosen for their ease of implementation and versatile nature. In this study, a correction method by modifying the position measurement in real time is introduced and compared to compensation tables. The reaction response of the numerical controller (NC) to the modification of its position feedback is studied, and a 5-axis machining experiment to validate the proposed solution is performed. The principle of the experiment is to impose a virtual volumetric error in the workspace by modifying a machining program, then to test separately the ability of compensation tables and the proposed method to correct the chosen virtual geometric errors. The aim is to obtain a corrected workpiece similar to the one machined with a nominal program. In this way, it is not necessary to identify the geometric errors of the machine’s structure to test the performance of software compensationmethods. The machined workpieces feature geometries that are easy to control, but the tool paths generated to produce them were complex enough to challenge the compensation methods. The ability of the proposed solution to correct the virtual volumetric error introduced by a modified machining program is evaluated at 98%. Indeed, roundness measurements show that over 99% of the added error has been corrected, with residuals lower than 5 μm. Furthermore, the joint trajectories monitored during machining are studied through a contouring error estimation. Nominal and compensated trajectories are 98% similar with the proposed solution, compared with 35% for compensation tables.

The presence of heterogeneities and significant property mismatches in soft composites lead to complex be­ haviors that are challenging to model with conventional analytical or numerical homogenization techniques. The present work introduces a micromechanics-informed deep learning framework to characterize microscopic dis­placements and stress fields in soft composites with periodic microstructures undergoing finite deformation. The main obstacle we address is the construction of specific loss functions incorporating intricate knowledge of finite strain homogenization theory, which is valid for arbitrary macroscopic deformation gradients. Notably, a multi-network model is utilized to describe the discontinuities in material properties and solution fields within the composites. These neural networks communicate with each other through interface traction and displacement continuity conditions within the loss function. In addition, to exactly impose the periodicity boundary in hex­agonal and square unit cells, the neural network architectures are modified by incorporating a number of trainable harmonic functions. A significant advantage of the current framework is that it allows for a straight­ forward solution of the governing partial differential equations expressed in terms of the first Piola-Kirchhoff stresses, eliminating the need for iterative formulations of the residual vector and tangent matrix required by classical numerical methods. We extensively assess the effectiveness of the proposed approach upon extensive comparison with isogeometric analysis to determine the displacement and Cauchy stress fields in square and hexagonal arrays of fibers/porosities, demonstrating neural networks as a powerful alternative to the conven­tional numerical approaches in finite deformation analysis of microstructural materials.