Portail HAL IMT Mines Albi
Not a member yet
5440 research outputs found
Sort by
Behavior-based multi-sensor navigation in an unknown environment
No full textInternational audienceThe paper presents a behavior-based, multi-sensor reactive navigation strategy for moving in unknown environments with complex-shaped obstacles or narrow passages. The obstacle avoidance includes an anticipation step to facilitate switching between navigation and avoidance reactive controllers. The first contribution is the design of a prediction algorithm that secures transitions between them. The second one, more general, lies in modeling the avoidance task by incorporating the anticipation within the same spiral-based motion framework. This modeling approach enables the design of a single stand-alone reactive controller for collision avoidance. Simulations based on ROS/Gazebo validate the approach, demonstrating its effectiveness
Effect of process parameters on surface integrity in laser powder bed fusion of Ti-6Al-4V alloy
No full textInternational audienceLaser powder bed fusion (LPBF) printing of Ti-6Al-4V often suffers from surface defects such as weld traces, unmelted particles, spatter, and porosity that degrade surface integrity and induce significant residual stresses due to a higher thermal gradient. This study aims to elucidate the evolution of surface topography and residual stresses as a function of build height and process energy input during LPBF fabrication. Ti-6Al-4V samples were produced at volumetric energy densities ranging from 21.8 to 65.4 J·mm⁻³ by varying laser power and scanning speed. Focus variation microscopy measurements show propagation of surface irregularities extending over several tens of layers. Higher energy densities resulted in stable melt pools and reduced surface roughness. The computed and measured residual stresses agreed well for the sound builds. In contrast, notable deviation was observed for the conditions that exhibited high porosity. Overall, the findings show that residual stress evolution in LPBF significantly influences surface morphology, which further supports the importance of process energy optimization to achieve a sound part
Structuring Federated Data Interoperability: A Multi-Level Framework
No full textInternational audienceCollaboration between organizations is essential for tackling global challenges, optimizing resources, and developing services. It depends on efficient communication to ensure smooth inter- and intra-organizational interactions. While existing frameworks define how to build and operate services, compatibility barriers persist. In this context, interoperability is crucial in ensuring a shared understanding among partners. However, significant challenges remain, as existing solutions often struggle to address the diverse concerns related to data, services, processes, and business operations. Traditionally, interoperability has relied on integrated and unified solutions with clearly defined requirements. However, emerging technologies are shifting toward a more federated approach, offering greater flexibility, automation, and agility. Yet, this approach remains loosely defined, with ambiguous requirements. This paper introduces a contextualization model that structures federated data interoperability into distinct levels, each representing a different degree of compliance with its definition. The model provides a standardized framework for guiding future implementations of federated data interoperability, focusing on middleware solutions that exchange, process, and synchronize data in collaborations. Furthermore, it addresses technical and conceptual barriers, and provides structured requirements for the concept. Finally, this work evaluates state-of-the-art federated interoperability solutions, positioning them within the defined levels and assessing their capabilities against specified requirements
Minimizing makespan in a multi-stage hybrid flow shop scheduling problem with dedicated machines
No full textInternational audienceThis paper addresses the scheduling problem for flow shops with multiple dedicated machines, commonly referred to as the Hybrid Flow Shop Scheduling Problem with Dedicated Machines (HFSSPDM). It corresponds to the case where dedicated machines are assigned to particular job types. The primary objective is to determine a schedule that minimizes the makespan, ensuring efficient utilization of machines. To tackle this challenging problem, we propose a Genetic Algorithm (GA), which is well-suited for solving complex scheduling problems. A preliminary analysis was conducted to evaluate the performance of the proposed GA in terms of solution quality and computation time across a set of HFSSPDM instances. The results demonstrate that the GA is capable of producing high-quality feasible solutions for several problem instances. Future efforts will focus on addressing the observed limitations to develop a more robust and efficient version of the Genetic Algorithm model
Experimental analysis and modeling of an acrylic thermoplastic resin polymerization: Influence of reactive mixture composition on temperature dependent kinetics
No full textInternational audienceManaging the in-situ polymerization of reactive thermoplastic mixtures represents a key challenge for themanufacture of thick composite parts. In particular, the highly exothermic free-radical polymerization of methylmethacrylate (MMA) exhibits complex kinetic phenomena, including the Trommsdorff effect, which arises fromdiffusional limitations. In this work, the reaction kinetics of Elium® C195E reactive mixtures were investigatedby differential scanning calorimetry (DSC) under various isothermal and non-isothermal conditions. Power-compensated DSC and low-mass samples ensured minimal temperature rise caused by the exothermic natureof the reaction, thus enabling the polymerization to be analyzed from a purely kinetic angle. By varying thenature and initial quantity of peroxide initiators, the influence of the available active chain amount on theoccurrence and intensity of the various polymerization phases could be highlighted. With a view to processingoptimization, a time and temperature dependent kinetic model was then developed to simulate the reaction rateof this acrylic resin initiated by 3 different organic peroxides. Based on the reaction scheme, this semi-empiricalmodeling approach yields highly accurate predicting results and enables to account for the occurring phenomenaduring MMA polymerization in isothermal ([343–383] K) and dynamic conditions ([1–20] K. min 1). This modeltherefore provides valuable support for the manufacture of Elium® matrix composites and thus limits issuesresulting from medium overheating
Tri‐reforming of Methane over a Hydroxyapatite‐Supported Nickel Catalyst Prepared by Cation Exchange
No full textInternational audienceCatalytic steam reforming of methane is a major process to produce synthetic gas (syngas), deployed at a large industrial scale. However, this is an energy‐intensive process since it is carried out at a high temperature above 900 °C with a large excess of water to preserve catalyst stability. Tri‐reforming of methane (TRM) is an alternative solution to steam reforming of methane, which allows valorizing not only methane and water but also carbon dioxide and oxygen, which are usually present in feedstocks such as biogas, landfill gas, or flue gas. Developing a highly performing catalyst for TRM under severe reaction conditions, in particular, with the inlet feeding composition close to the chemical stoichiometry, is meaningful to optimize process energy efficiency. In this work, a hydroxyapatite‐supported nickel catalyst containing 1 wt% Ni is synthesized by cation exchange and investigated in TRM. High activity and very good stability are obtained, making it among the most‐performing catalysts in TRM, in comparison with the data from the literature
Évaluation de la robustesse des modèles de prévision des séries temporelles sous perturbations
No full textNational audienc
Modèles de connaissances pour la recommandation des services métiers personnalisés aux besoins des usagers : application aux services de prévention de perte d'autonomie
No full textThe prevention of loss of autonomy in the elderly has become a key priority for healthcare systems and social organizations, in response to the challenges posed by a rapidly aging population. The unique characteristics and specific needs of this vulnerable group require personalized interventions, not only to improve their quality of life, but also to optimize available institutional resources. However, implementing such tailored solutions presents both technical and scientific challenges, particularly regarding the representation and use of complex knowledge. In this context, this thesis aims to develop a socio-cyber-physical environment to support the design of personalized intervention plans that address the specific needs of elderly individuals and contribute to the proactive management of their frailty. To achieve this, it is essential to formalize knowledge that enables the characterization of both users and relevant service offerings. A structured representation of this information ensures the alignment between individual needs and appropriate services, while integrating diverse data sources, including frailty assessments and accessible service repositories. Our approach to modeling user needs draws upon several standardized assessment tools. These indicators enable the construction of detailed user profiles, which the system uses to generate service recommendations. Candidate services are retrieved from repositories such as the french National Directory of Health and Social Care Services and Support (ROR), and further enriched with additional data sources to better align with user preferences. The project goes beyond simply modeling user profiles and services. It includes an in-depth analysis of how these knowledge representations can be leveraged to generate relevant and personalized recommendations. We investigate the most suitable recommendation techniques for this context, exploring how the knowledge can be integrated and interlinked within a recommendation system. Widely-used recommendation techniques are examined and selected based on their applicability and effectiveness in a domain where personalization and reliability are critical. One of the innovative aspects of this work is the explainability of the recommendations: each proposed service is explicitly justified by the specific needs identified for each user, ensuring transparency in the matching process. This explainability is based on the concept of operational activities, which link each service to well-defined needs of elderly individuals. Finally, the effectiveness of the recommendation system is evaluated through a series of use cases based on profiles of elderly users. This evaluation, conducted in a rigorous and iterative manner, allows for continuous refinement of the system and optimization of the relevance of the recommendations. The results demonstrate a promising alignment between the suggested services and the users' needs, while also accounting for their individual preferences. This thesis lies at the intersection of industrial engineering and computer science, with the ambition of providing practical tools to healthcare professionals and institutions. It aims to support the care of elderly individuals and promote more efficient management of institutional resources. The outcomes of this project offer meaningful contributions to enhancing the quality of life for older adults and reducing the organizational costs associated with their care.La prévention de la perte d'autonomie chez les personnes âgées est devenue une priorité pour les systèmes de santé et les organismes sociaux, face aux défis imposés par le vieillissement rapide de la population. Les caractéristiques et les besoins spécifiques de cette population vulnérable requièrent des interventions personnalisées, autant pour améliorer leur qualité de vie que pour optimiser les ressources institutionnelles disponibles. Cependant, la mise en oeuvre de solutions personnalisées rencontre des défis techniques et scientifiques, particulièrement en ce qui concerne la représentation et l'exploitation de connaissances complexes. Dans ce contexte, cette thèse vise à développer un environnement socio-cyber physique pour faciliter la conception de plans d'intervention personnalisés, répondant aux besoins spécifiques des personnes âgées et contribuant ainsi à la gestion proactive de leur fragilité. Pour y parvenir, il est essentiel de formaliser les connaissances permettant de caractériser à la fois les usagers et les services métiers pertinents. Une représentation structurée de ces informations assure la connexion entre les besoins des personnes âgées et les services pouvant y répondre, tout en intégrant des données variées, notamment des évaluations de fragilité et des bases de services accessibles. Dans notre approche, la modélisation des besoins des usagers s'inspire de plusieurs outils d'évaluation standardisés. Ces indicateurs permettent de dresser un profil détaillé des besoins individuels, que le système utilise pour générer des recommandations de services. Les services candidats à la recommandation sont quant à eux extraits de répertoires tels que le Répertoire Opérationnel des Ressources (ROR), enrichis par des bases de données complémentaires pour répondre de manière encore plus fine aux préférences des utilisateurs. Ce projet dépasse la simple modélisation des profils et services en analysant l'utilisation approfondie de ces connaissances pour générer des recommandations pertinentes et adaptées. Il examine comment intégrer et interconnecter ces données au sein d'un système. Les techniques de recommandation les plus appropriées pour ce contexte sont ainsi étudiées afin de justifier les choix opérés, en fonction de leur efficacité dans un domaine où la personnalisation et la fiabilité des recommandations sont essentielles. L'un des aspects novateurs de ce travail est l'explicabilité des recommandations : chaque service proposé est justifié par les besoins identifiés pour chaque utilisateur. Cette explicabilité repose sur la notion d'activités opérationnelles, permettant de relier chaque service aux besoins spécifiques des personnes âgées. Enfin, l'efficacité du système de recommandation est évaluée à travers une série de cas d'usage basés sur des profils de personnes âgées. Cette évaluation, menée de manière rigoureuse et itérative, permet d'affiner continuellement le système et d'optimiser la pertinence des recommandations. Les résultats montrent une adéquation prometteuse entre les services suggérés et les besoins des usagers, tout en prenant en compte leurs préférences individuelles. Ainsi, cette thèse se situe à l'intersection du génie industriel et de l'informatique, avec pour ambition de fournir des outils concrets aux professionnels de santé et aux institutions, facilitant la prise en charge des personnes âgées et contribuant à une gestion plus efficiente des ressources. Les perspectives offertes par ce projet peuvent apporter une contribution significative à l'amélioration de la qualité de vie des personnes âgées et à la réduction des coûts organisationnels liés à leur prise en charge
Thermal field estimation in CFRTP composites using an attention-enhanced U-Net
No full textInternational audienceThis study presents a surrogate model based on the convolutional U-Net architecture to predict the thermal field in a carbon fibre-reinforced thermoplastic tape at the microscale during brief and localized heating. Leveraging microstructure data within a machine learning framework, the proposed model aims to enhance the accuracy of temperature field predictions at a low computational cost. The incorporation of a co-attention mechanism to handle image channels of different nature significantly improves precision, resulting in a strong correlation between the model’s predictions and the ground truth obtained from the numerical solution of the heat equation. This capability enables rapid assessment of diverse microstructures, facilitating optimization and real-time applications in manufacturing settings
Synthesis and characterization of gelatin-hydroxyapatite nanohybrid composites
No full textInternational audienceGelatin-based nanocomposites have gained significant attention as sustainable and multifunctional biomaterials due to their natural origin, biocompatibility, and tunable properties, making them strong candidates for advanced structural applications, such as matrix fillers and functional coatings in porous substrates like leather. But, pure gelatin suffers from limited mechanical strength, thermal instability, and rapid solubility in aqueous environments, necessitating structural reinforcement. Nano-calcium hydroxyapatite (CaP) is a well-established inorganic additive that enhances the mechanical, thermal, and structural performance of gelatin matrices; however, its tendency to agglomerate in gelatin limits uniform dispersion, and consequently restricts its reinforcement capabilities. This study focuses on the synthesis and characterization of a novel nanohybrid composite, integrating gelatin nano-emulsion as the biopolymer matrix with different weight ratios of nano-CaP colloidal suspension (CaP10 % and CaP20 %) as the reinforcing phase. Various characterization techniques were used to evaluate the influence of nano-CaP on interfacial interaction and effect on microstructure. Fourier transform infrared spectroscopy and X-ray diffraction results show calcium hydroxyapatite as the main phase (75 %) in the nanohybrid composite while scanning electron microscope confirmed its intimate contact and homogenous dispersion. The 10 % CaP loading provided a better physical and thermo response, including a higher glass transition temperature (220 °C), and increased hydrophobicity (45 % increase in water contact angle). These enhancements in properties are attributed to the homogenous dispersion of nano-CaP in the nanogelatin matrix and the strong interfacial interaction between the particle surfaces and the nanogelatin molecule. Overall, the results demonstrate the potential of colloidal nano-CaP suspension as a dual-function agent for cross-linking and thermo-mechanical reinforcement in gelatin-based nanocomposites