Portail HAL IMT Mines Albi
Not a member yet
5440 research outputs found
Sort by
Tri-reforming of methane over a highly-dispersed hydroxyapatite-supported nickel catalyst
No full textInternational audienceTri-reforming of methane is a promising process to produce synthetic gas via the reforming of methane simultaneously with three different oxidants: water vapor, carbon dioxide and oxygen. The thermodynamic aspect of this reaction requires high reaction temperatures (above 700 °C) to reach high methane conversion, while the kinetic aspect requires a solid nickel-based catalyst to reach high reaction rates. Consequently, the solid catalyst used in TRM is exposed to several risks of deactivation, and the development of a performing catalyst is a great challenge for TRM process. In this work, it has been demonstrated that hydroxyapatite (HAP) supported nickel catalyst prepared by ion exchange / ultrasonic vibration led to the formation of nickel nanoparticles and sub-nanoparticles on the HAP support surface, and the resulting catalyst showed very good catalytic performance
Fischer-Tropsch synthesis on carbon-supported Co catalysts coated on metallic foams
No full textInternational audienceThe Fischer-Tropsch synthesis (FTS) is a highly exothermic reaction, widely applied to convert syngas to liquid fuels. Temperature control is particularly critical in order to ensure longevity of the catalyst, thermo-mechanical reliability of the entire process and to optimize the product selectivity. Open-cell foams have attracted increasing attention for the intensification of the FTS, thanks to their good thermal conductivity. However, to-date, FTS is limited to classical oxide-supported catalysts. Here, new FTS catalysts have been developed by dip-coating powder Co/C catalyst onto a metallic foam. The beneficial effect of using metallic foams is evidenced, notably by a higher CTY and SC5+, but also by a better stability of the structured catalysts as compared to the corresponding powder catalysts. The impact of the carbon structure on the catalyst performance is also highlighted
Healthcare facilities spatial layout design: a review on case studies
No full textInternational audienceThe spatial layout design of healthcare facilities plays a critical role in ensuring the efficiency of patient flow, staff movement, and material logistics. As hospitals face increasing demand and overcrowding, the need for adaptable and well-planned configurations becomes ever more pressing. This review focuses on studies that have applied various methodologies to real-world case studies within this context. Although significant advancements have been made in flow analysis, wayfinding systems, and spatial layout planning, several critical gaps persist. These include the limited integration of emerging technologies into the design process, the absence of holistic, systematic frameworks for designing optimally functioning hospitals from inception, and insufficient interdisciplinary collaboration among architects, healthcare administrators, and end-users like medical professionals and patients. The findings of this review point toward future research opportunities aimed at closing these gaps, while also emphasizing the importance of further comparative studies to understand discrepancies between theoretical models and practical applications
Étude et définition d’une approche outillée dédiée à la modélisation et au suivi des portefeuilles de projets en environnement instable
No full textThe doctoral project concerns the study and definition of an innovative approach to modelling and monitoring project portfolios in a potentially multi-partner context. The initial premise is that “instability is now the norm”. In this context, the main issues of the thesis concern the identification, assessment and mitigation of significant disruptions to project portfolios and their management. Each of these three levels is integrated into a so-called “socio-technical” environment (mixing social and psychological considerations with more technical and technological aspects), constituting a source of complexity and richness for the work envisaged. The first level of identification mainly concerns linking data and information flows relating to the projects under consideration with formal elements of project modelling. As for the second level of assessment, the challenge will mainly concern the ability to exploit dynamic models of projects and their environments to anticipate the impact of potential events (risks and opportunities, endogenous or exogenous, as identified during the identification phase). Finally, the third level of mitigation will be based on evaluating the cross or joint impacts of potential risks and opportunities to guide users towards the most relevant and profitable decision combinations for managing the project portfolios under consideration. This work will continue the research activities carried out since 2018 on the POD (Physics of Decision) theme.Le projet doctoral concerne l'étude et la définition d'une approche innovante de modélisation et de suivi des portefeuilles de projets en contexte potentiellement multipartenaires. Le postulat initial est le suivant : « l'instabilité est désormais la norme ». Dans ce contexte, les enjeux principaux du sujet de thèse concernent l'identification, l'évaluation et la réduction des perturbations significatives que peuvent rencontrer les portefeuilles de projets et leur pilotage. Chacun de ces trois niveaux s'intègre dans un environnement dit « socio-technique » (mixant à la fois des considérations sociales et psychologiques avec des aspects plus techniques et technologiques) qui constitue une source de complexité et de richesse pour les travaux envisagés. Le premier niveau de l'identification, la vision initiale relève principalement de la mise en relation de flux de données et d'informations relatives aux projets considérées avec des éléments formels de modélisation des projets. Le deuxième niveau de l'évaluation, l'enjeu concernera principalement la capacité d'exploiter les modèles dynamiques des projets et de leurs environnements afin d'anticiper les impacts d'événements potentiels (risques et opportunités, endogènes ou exogènes, tels que recensés lors de la phase d'identification). Le troisième niveau de la réduction, s'appuiera sur l'évaluation des impacts croisés ou conjoints des risques et opportunités potentiels afin de guider les utilisateurs vers les combinaisons de décisions les plus pertinentes et profitables pour la gestion des portfolios de projets considérés. Il est à noter que ces travaux s'inscriront dans la continuité des activités de recherche menées depuis 2018 sur la thématique POD (Physics of Decision)
Automatic visual inspection of mechanical assemblies via 3D point cloud classification with deep neural networks
No full textIn this work, we are focused on conformity control of complex aeronautical mechanical assemblies,typically an aircraft engine, at the end or in the middle of the assembly process. A 3D scannercarried by a robot arm provides acquisitions of 3D point clouds which are further processed by deepclassification neural networks. The Computer-Aided Design (CAD) model of the inspected mechanicalassembly is available, and the proposed approach relies on it. The models are trained on syntheticdata, generated from the CAD models
Addressing viscosity-driven singularities: accurate development of thermo-elasto-visco-plastic constitutive models
No full textInternational audienceA novel analytical-mathematical formulation for the multi-physics thermo-elasto-visco-plastic (TEVP) behavior of materials with nonlinear combined hardening is proposed. New closed-form expressions for the incremental visco-plastic multiplier (IVPM) and the consistent tangent operator (CTO) were derived. Specifically, all stiffness, hardening, and viscous coefficients were treated as temperature-dependent, and their temperature derivatives were explicitly included in the analytical solution. A UMAT (User Material) subroutine was programmed and implemented to compute the IVPM, CTO, and isotropic, kinematic, and viscous stresses for TEVP modeling. Finite element (FE) models were created and compared for the Abaqus® built-in material model and the developed UMAT subroutine. The IVPM and CTO equations were successfully validated and the influence of the initial IVPM value on the accuracy of the results and the run time of simulations was examined for the first time. It was found that, in the Newton-Raphson method, the initial IVPM value must not only be nonzero to avoid singularity issues, but also be less than or equal to 10−8 to ensure accurate results. In addition, the initial IVPM value did not influence computational efficiency. Ultimately, based on a comparative study of analytical solutions, UMAT-driven simulations, and standard Abaqus simulations, the developed formulation enables accurate prediction of strains, stresses, and temperatures in TEVP problems, providing a solid foundation for modeling industrial manufacturing processes such as quenchin
Predicting powder flows in convective mixers through multi-scale rheology
No full textInternational audienceThe analysis of powder flow induced by an agitation device requires further investigation to optimise the design and operation of industrial mixers. Systems equipped with an agitation mobile, like convective paddle mixers, are primarily described by the torque resulting from the interaction between the paddles and the powder bed. Several studies have identified, through dimensional analysis, correlations between a torque-related parameter, the power number, and an agitation speed-related parameter, the Froude number [1], [2]. However, these correlations often neglect the characteristics of the powder being agitated. Therefore, our study aims to incorporate these characteristics into a correlation between two dimensionless numbers: the effective friction coefficient μeff and the inertial number I.Our research involves mixers with capacities ranging from 0.3 l to 3000 l and diverse configurations, including variations in the number of paddles, the number of shafts and their orientation—horizontal and vertical. The flow analysis focuses on the mesoscale, corresponding to the shear bands formed on either side of the paddles [3]. Experimental results indicate that the flow occurs in the dense regime [4], [5] and that the developed μ(I)-rheology law can be transposed from one mixer to another, regardless of configuration. Moreover, μ(I)-rheology accounts for particle size differences for powders with similar particle shapes, rather better for deep powder beds than shallow ones [6].This rheological law links macroscopic characteristics (e.g., torque, mixer dimensions, operating conditions) to the microscopic characteristics of the powder. This approach enables the description of powder behaviour across different regimes and facilitates scale-up. Bibliography[1]L. Legoix, C. Gatumel, M. Milhé, and H. Berthiaux, ‘Analysis of powder flow and in-system rheology in a horizontal convective mixer with reclining blades’, Particulate Science and Technology, vol. 36, no. 8, pp. 955–966, Nov. 2018, doi: 10.1080/02726351.2017.1331284.[2]M. Sato, K. Miyanami, and T. Yano, ‘Power reqiurement of horizontal cylindrical mixer’, Powder Technology, vol. 16, pp. 3–7, 1979.[3]J. Lehuen, J.-Y. Delenne, A. Duri, and T. Ruiz, ‘Forces and flow induced by a moving intruder in a granular packing: coarse-graining and DEM simulations versus experiments’, Granular Matter, vol. 22, no. 4, p. 78, Nov. 2020, doi: 10.1007/s10035-020-01047-5.[4]F. Da Cruz, ‘Friction and jamming in dry granular flows’, Ecole des Ponts ParisTech, 2004.[5]GDR MiDi, ‘On dense granular flows’, Eur. Phys. J. E, vol. 14, no. 4, pp. 341–365, Aug. 2004, doi: 10.1140/epje/i2003-10153-0.[6]H. Boussoffara, C. Gatumel, B. Malécot, M. Viau, and H. Berthiaux, ‘A rheological law to describe powder agitation in a lab-scale paddle mixer: Shear band observation and dimensional analysis’, Powder Technology, vol. 451, p. 120469, Feb. 2025, doi: 10.1016/j.powtec.2024.120469
Improved human activity recognition through controllable GAN-Generated synthetic data and large Language models for classification
No full textInternational audienceHuman Activity Recognition (HAR) plays a critical role in healthcare monitoring and smart home systems, enabling tracking of patient movements, fall detection, and daily activity monitoring. However, HAR faces challenges due to the scarcity of diverse, large-scale datasets and the absence of sufficient abnormal activity samples necessary for detecting rare but critical health events. This paper addresses these challenges through advanced synthetic data generation and state-of-the-art classification techniques. We introduce a Generative Adversarial Network (GAN) for time-series data to generate synthetic samples, significantly expanding the WISDM dataset and incorporating an ‘abnormal’ activity class to enhance dataset diversity and real-world applicability. The fidelity of the synthetic data is rigorously evaluated using Dynamic Time Warping (DTW), achieving an average distance of 56.1, demonstrating strong alignment with real data distributions. For classification, we leverage transformer-based models, which have shown superior performance over traditional HAR methods such as CNNs and LSTMs. Our approach achieves 91.4% accuracy and an 87.6% macro F1-score, surpassing state-of-the-art methods that report accuracy in the range of 85–89% and macro F1-scores of 81–85%. These results highlight the effectiveness of integrating Controllable GAN-generated synthetic data with LLM-based classification, significantly improving recognition of rare activities by 15% compared to SOTA benchmarks. This work contributes to HAR by providing a framework for dataset enhancement and classification, paving the way for more robust and adaptable activity recognition systems, particularly in data-scarce environments. The implications for healthcare are substantial, with the potential to enhance patient monitoring, improve early detection of critical health events, and enable more efficient healthcare delivery
Effect of Structural Defects on Electrical Conductivity of Graphitic Biocarbon
No full textInternational audienceThe catalytic graphitization process is considered a promising method for producing well-structured advanced biocarbon with a higher degree of graphitization at lower temperatures. The challenge with advanced biocarbon materials lies in determining their properties and expanding their applications beyond standard uses in soil, the environment, and fuel to more advanced areas, such as composites, electronics, photonics, and energy storage. Data on their properties remains scarce. This study examines the nanostructure chemistry of graphitic biocarbon produced through the catalytic graphitization of cellulose at various temperatures, accompanied by a comprehensive investigation of defects. The new approach developed uses the iron−calcium bimetallic catalytic graphitization. The conductive grains of graphitic biocarbon become active, improving the electrical conductivity to approximately 10−2 S m−1, which strongly depends on defects from grain boundaries and vacancies. Furthermore, the conductive capacity of biocarbon is enhanced by an increase in graphitic sp2 carbon content and a decrease in defect concentration, such as carbon vacancy defects, which decline from 11.77% to 6.57% between 1200 and 1800 °C. Herein, we assert that high-temperature catalytic graphitization generates defective graphitic biocarbon, and these defects significantly influence its properties, establishing it as a potential semiconductive material that expands its applications in photonics and electro-optics, thus opening new opportunities
Evaluating the Robustness of Time Series Forecast Models Under Disruptions
No full textInternational audienceIn uncertain, hyper-connected and fast-changing environments, supply chains are becoming more prone to disruptions and unexpected events which affect their stability and ultimately the future consumer demand. Traditional time series forecasting approaches fail to take into consideration environmental uncertainties related to the volatile nature of the supply chains. This paper provides an evaluation of the robustness of time series forecast models under disruptions and future uncertainties. By introducing disruption scenarios with varying intensities and durations, the study evaluates the robustness of forecast models using a range of statistical and machine learning models. This study was evaluated on M3 monthly data to test the performance of forecast models and selection strategies using various accuracy metrics. The results underscore the need for a new robust forecast model selection approach to find a trade-of between accuracy and robustness to future uncertainties