HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
CotaSigne, une expérience de recherche à plusieurs mains
No full textInternational audienceCotaSigne is a linguistic research project aimed at collecting, analysing and evaluating the production of narratives in French Sign Language (LSF) of children aged between 4 and 11 years. Meeting the shared needs of linguistic research in language acquisition, language therapy and teaching in/of LSF, this research project is participatory in nature, involving researchers, educational professionals and language therapists, deaf and hearing, and includes an online platform with computerised scoring and visualisation of results. In this article, we highlight the way in which CotaSigne fits in with the scientific and deontological heritage of Patrice Dalle and his work, by relying on uses and users on the one hand, by associating the different expertises to the different stages of the research process on the other hand, and by putting LSF at the heart of the reflection, both as the language studied and as the working language.CotaSigne est un projet de recherche linguistique visant le recueil, l’analyse et l’évaluation de production de récits en LSF, chez des enfants âgés de 4 et 11 ans. Répondant aux besoins partagés de la recherche linguistique en acquisition du langage, de la prise en charge orthophonique et de l’enseignement de /en LSF, le projet se décline de manière participative, associant des chercheuses, enseignantes, orthophonistes, sourdes et entendantes, et comprend une plateforme en ligne avec une cotation et une visualisation des résultats informatisées. Dans cet article, nous soulignons la façon dont CotaSigne s’inscrit dans l’héritage scientifique et déontologique de Patrice Dalle et de ses travaux, en s’appuyant sur les usages et les usagers d’une part, en associant les différentes expertises aux différentes étapes du processus de recherche d’autre part, et en mettant la LSF au coeur de la réflexion, tant comme langue étudiée que comme langue de travail
Analyse de forme pour la Rétro-Ingénierie de modèles 3D : application à l'industrie aéronautique
No full textThis thesis addresses the challenges related to digital continuity of companies by focusing on Shape Analysis (SA) activities for Computer-Aided Design (CAD) model Reverse Engineering (RE) processes. Activities such as comparison, retrieval, and identification of CAD models from databases are studied to enable product definition information to be retrieved and inferred. In the aerospace sector, data volume and complexity of aerodynamic surfaces CAD models are a significant challenge. The concept of dissimilarity, commonly used in SA activities, fails to accurately distinguish between 3D shapes that are very similar but not identical, thus compromising the reliability of the information obtained. To address this issue, a new 3D shape descriptor, the AISM (Algebraic Implicit Surface Models), is proposed. Based on advanced techniques of computational geometry, statistical analysis and machine learning, it adapts to the specific geometric features of complex aerodynamic surfaces, and integrates with existing SA methods, such as matching, retrieval, and classification. A new activity for the precise identification of shapes called labeling is also introduced to enhance and reinforce the search and exploration capabilities within CAD model databases. In addition, the Knowledge-Based Reverse Engineering (KBRE) model is introduced. This data model integrates advanced features for processing and analyzing CAD model databases, to facilitate the development, implementation and experimentations on advanced SA techniques into RE processes.Cette thèse aborde les défis liés à la continuité numérique des entreprises en se concentrant sur l'exploitation des activités d'Analyse de Formes 3D (SA) dans les processus de Rétro-Ingénierie (RE) des modèles CAO. Les activités telles que la comparaison, la recherche et l’identification de modèles CAO dans les bases de données sont étudiées pour permettre de recréer et retrouver des informations relatives à la définition des produits. Dans le secteur aéronautique, la volumétrie et la complexité des modèles CAO de surfaces aérodynamiques posent un défi majeur. La notion de dissimilarité généralement utilisée dans les activités de SA ne parvient pas à distinguer précisément des formes 3D très proches mais non identiques, ce qui compromet la fiabilité des informations obtenues. Pour répondre à ce problème, un nouveau descripteur de forme 3D, les AISM (Algebraic Implicit Surface Models), est proposé. Sur la base de techniques avancées de géométrie computationnelle, d'analyse statistique et d'apprentissage automatique, il s'adapte aux spécificités géométriques des surfaces complexes, et s'intègre aux méthodes courantes de SA, comme le matching, le retrieval, et la classification. De plus, une nouvelle activité d'identification précise de formes, le labeling, est proposée pour améliorer et renforcer les capacités de recherche et d'exploration dans les bases de modèles CAO. Par la suite, le Knowledge-Based Reverse Engineering (KBRE) Model est introduit. Ce modèle de données intègre des fonctionnalités avancées de traitement et d’analyse de bases de modèles CAO, afin de faciliter le développement, l’implémentation et l’expérimentation de techniques avancées de SA dans les processus de RE
1st International Conference on Synergy between Multi-physics/Multi-scale Modeling and Machine Learning,
No full textInternational audienceS4ML 2024 is the first in the series of International Conferences focusing upon Synergybetween Multiphysics/ Multiscale Modeling and Machine Learning, to be held in different countries in Central Europe. These meetings will be organized under umbrella of CEACM, ‘Central European Association for Computational Mechanics’, regrouping (in alphabetic order) the following Central European countries: Austria, Bosnia-Herzegovina, Croatia, Czech Republic, Hungary, Slovakia and Slovenia. For this particular Conference, we have additional sponsors, ANUBiH, ‘Academy of Sciences and Arts of Bosnia and Herzegovina’ and from France, IUF, ‘Institut Universitaire de France’. The institution organizing S4ML 2024 is the Faculty of Civil Engineering of Czech Technical University in Prague. The S4ML 2024 is held from 19 to 24 June 2023 in Prague, after the short course offered to young participants held from 17 to 18 June 2024. The Conference seeks to provide a platform for learning from some of the worlds' leading specialists in numerical methods, coming from different engineering disciplines and applied mathematics
Uniform in Number of Neighbor Consistency and Weak Convergence of k-Nearest Neighbor Single Index Conditional Processes and k-Nearest Neighbor Single Index Conditional U-Processes Involving Functional Mixing Data
No full textInternational audienceU-statistics are fundamental in modeling statistical measures that involve responses from multiple subjects. They generalize the concept of the empirical mean of a random variable X to include summations over each m-tuple of distinct observations of X. W. Stute introduced conditional U-statistics, extending the Nadaraya–Watson estimates for regression functions. Stute demonstrated their strong pointwise consistency with the conditional expectation r(m)(φ,t), defined as E[φ(Y1,…,Ym)|(X1,…,Xm)=t] for t∈Xm. This paper focuses on estimating functional single index (FSI) conditional U-processes for regular time series data. We propose a novel, automatic, and location-adaptive procedure for estimating these processes based on k-Nearest Neighbor (kNN) principles. Our asymptotic analysis includes data-driven neighbor selection, making the method highly practical. The local nature of the kNN approach improves predictive power compared to traditional kernel estimates. Additionally, we establish new uniform results in bandwidth selection for kernel estimates in FSI conditional U-processes, including almost complete convergence rates and weak convergence under general conditions. These results apply to both bounded and unbounded function classes, satisfying certain moment conditions, and are proven under standard Vapnik–Chervonenkis structural conditions and mild model assumptions. Furthermore, we demonstrate uniform consistency for the nonparametric inverse probability of censoring weighted (I.P.C.W.) estimators of the regression function under random censorship. This result is independently valuable and has potential applications in areas such as set-indexed conditional U-statistics, the Kendall rank correlation coefficient, and discrimination problems
Valorisation de la lignine de drêches de brasserie : fractionnement, purification et couplages diazoïques pour des applications en antimicrobiens
No full textBeer, one of the most widely consumed fermented beverages globally, generates significant by-products, primarily brewer's spent grains (BSG). In France, the annual production of dried BSG is estimated at approximately 0.7 million tons. This by-product is characterized by its high moisture content (around 80%) and a rich composition of proteins, minerals, and biopolymers, including lignin, cellulose, and hemicelluloses. After protein extraction, the main objective of this study was to investigate effective methods for lignin extraction using the Organosolv (AVIDEL) process, followed by an assessment of its chemical valorization through diazo coupling reactions for potential antimicrobial applications. The first phase of the research focused on the extraction of proteins from BSG. Two extraction techniques were evaluated: (i) a ball milling approach with water as the solvent during the preliminary screening, followed by optimization using water and sodium dodecyl sulfate (SDS) solution, and (ii) an ultrasonic-assisted extraction method, comparing water and an alkaline solvent (110 mM NaOH). The ultrasonic method with the alkaline solvent proved to be the most efficient, achieving a protein extraction yield of 71.83%±0.65 under optimal conditions (250 W, 60% cycle, 20 minutes, 20:1 liquid-to-solid ratio). Protein quantification was performed using elemental analysis. The second focused on lignin extraction from the deproteinized BSG. Microwave-assisted fractionation was optimized, resulting in a lignin yield of 50.28±1.14% and a purity of 85%. This yield was obtained using a solvent mixture of acetic acid, formic acid, and water in a 40/40/20 (%v/v) ratio, with a liquid/solid ratio of 20 and an extraction time of 50 minutes. FTIR spectroscopy identified key functional groups within the lignin structure. In addition, 2D-HSQC NMR analysis revealed significant signals in the aromatic region, corresponding to lignin's guaiacyl (G), syringyl (S), and p-hydroxyphenyl (H) units, as well as their inter-unit linkages. The final phase of this study involved the chemical valorization of lignin by functionalizing its phenolic groups through azobenzene grafting, known for its photochromic and antimicrobial properties. The degree of functionalization ranged from 17% to 49%, depending on the aniline derivatives used. Antimicrobial assays demonstrated that functionalized lignins exhibited significantly enhanced antibacterial activity compared to unmodified lignin, with up to 55% inhibition of E. coli bacterial growth at relatively low concentrations of the azobenzene-modified compounds. These findings demonstrate the potential of functionalized lignin derivatives for innovative antimicrobial applications, suggesting a promising pathway for the development of bio-based antimicrobial materials.La bière, l'une des boissons fermentées les plus consommées au monde, génère un sous-produit majeur : les drêches de brasserie. En France, la production annuelle de drêches est estimée à 0,7 million de tonnes. Ces résidus se caractérisent par une teneur élevée en humidité (environ 80 %) et une composition riche en protéines, minéraux et biopolymères, notamment la lignine, la cellulose et les hémicelluloses. Après l’extraction des protéines, l’objectif principal de cette étude est d'identifier la méthode la plus efficace pour l’extraction de la lignine, en utilisant le procédé Organosolv (AVIDEL) et d’étudier sa valorisation chimique via des couplages diazoïques pour des applications en tant qu’agent antimicrobien. La première partie de ce travail a porté sur l'extraction des protéines à partir des drêches de brasserie. Deux méthodes d'extraction ont été étudiées : l'utilisation d'un broyeur à billes avec de l'eau comme solvant pour un criblage initial, suivi de l'optimisation avec une solution d'eau + SDS (sodium dodécyl sulfate) et l'application d'ultrasons comparant un solvant aqueux à un milieu alcalin (NaOH 110 mM). La méthode utilisant les ultrasons avec un solvant alcalin s’est avérée la plus efficace, permettant d'obtenir un taux de récupération de protéines 71,8% ± 0,7 sous des conditions optimales : une puissance de 250 W, un cycle de 60 %, un temps de traitement de 20 minutes et un rapport liquide/solide de 1 :20. Ce rendement a été quantifié par analyse élémentaire. La deuxième partie de la recherche a porté sur l'optimisation de la délignification des drêches par micro-ondes, avec un taux de récupération de lignine de 50.28 %±1.14 et une pureté de 85% en masse. Ce rendement a été obtenu avec un mélange de solvants composé d'acide acétique, d'acide formique et d'eau dans des proportions de 40/40/20 (%v/v), un rapport liquide/solide de 20 et un temps d'extraction de 50 minute. Les analyses par spectroscopie infrarouge ont permis d’identifier des groupes fonctionnels caractéristiques de la lignine. De plus, l’analyse par RMN-HSQC (spectroscopie par résonance magnétique nucléaire- Heteronuclear Single Quantum Coherence) a révélé des signaux significatifs dans la région aromatique, indiquant la présence des unités monomères de lignine guaiacyl (G), syringyl (S) et p-hydroxyphényle (H), ainsi que leurs liaisons inter-unités. La troisième partie de ce travail s'est concentrée sur la valorisation chimique de la lignine via la fonctionnalisation de ses groupes phénoliques par greffage d'azobenzènes, connus pour leurs propriétés photochromiques et antimicrobiennes. Le degré de fonctionnalisation obtenu variait entre 17% et 49% en fonction des anilines utilisées. Les tests antimicrobiens ont révélé une activité accrue des lignines fonctionnalisées par rapport à la lignine non modifiée, avec une inhibition de la croissance d’E. coli atteignant jusqu’à 55% à de faibles concentrations d'azobenzènes. Ces résultats mettent en évidence le potentiel des lignines fonctionnalisées pour des applications antimicrobiennes innovantes
Analytical and numerical modelling of the effect of filter medium resistance on the mechanical dewatering of wastes and industrial filter cakes
No full textInternational audienceTea, one of the most popular beverages worldwide, generates a substantial amount of spent leaves, often directly discarded although they may still contain valuable compounds. This study aims to optimize the extraction of polyphenols from spent black tea (SBT) and spent green tea (SGT) leaves while also exploring their antioxidant and antibacterial properties. Response surface methodology was utilized to determine the optimal experimental conditions for extracting polyphenols from SBT and SGT. The total phenolic content (TPC) was quantified using the Folin–Ciocalteu method, while antioxidant activity was evaluated through the DPPH assay. Antibacterial activity was assessed using the disk diffusion method. Additionally, high-performance liquid chromatography (HPLC) was employed to analyze the phytochemical profiles of the SBT and SGT extracts. Optimal extraction for SBT achieved 404 mg GAE/g DM TPC and 51.5% DPPH inhibition at 93.64 °C, 79.9 min, and 59.4% ethanol–water. For SGT, conditions of 93.63 °C, 81.7 min, and 53.2% ethanol–water yielded 452 mg GAE/g DM TPC and 78.3% DPPH inhibition. Both tea extracts exhibited antibacterial activity against Gram-positive bacteria, with SGT showing greater efficacy against S. aureus and slightly better inhibition of B. subtilis compared to SBT. No activity was observed against the Gram-negative bacteria E. coli and S. typhimurium. HPLC analysis revealed hydroxybenzoic acid as the main phenolic compound in SBT (360.7 mg/L), while rutin was predominant in SGT (42.73 mg/L). The optimized phenolic-rich extracts of SBT and SGT demonstrated promising antioxidant and antibacterial potential, making them strong candidates for use as natural health boosters in food products
Text reuse as cultural practice: intertextuality in the 18th-century digital archive
No full textInternational audienceText reuse, encompassing direct citations, paraphrases and allusions, represents a key aspect of intertextuality – a concept central to literary theory since the 1960s. This paper highlights how computational methods, particularly automatic text-reuse detection, can illuminate the complex system of intertextual exchange that informs 18th-century literary culture, focusing on significant works like the Encyclopédie and Voltaire's correspondence. By employing advanced techniques such as sequence alignment and social network analysis, we uncover hidden patterns of influence, citation strategies and the subtle interplay between originality and imitation in Enlightenment literature. The paper also considers the implications of these findings for modern understandings of authorship, originality and textuality, drawing connections to contemporary digital humanities practices. The paper ultimately aims to recontextualise the Enlightenment as a period of intense intertextual productivity, where the reuse of texts was not merely a scholarly exercise but a dynamic and essential component of literary creation
Analyse numérique des propriétés thermoélastiques effectives des nanocomposites
No full textNational audienceLe comportement thermoélastique d’un nanocomposite constitué d’une matrice renforcés d’inclusions cylindrique est étudié grâce à un modèle éléments finis. Pour prendre en compte de l’effet de taille spécifique aux nanocomposites, une interface imparfaite cohérente est introduite entre la matrice et les inclusions. L’estimation des propriétés effectives permet de mettre en évidence l’influence de la taille des inclusions et des propriétés introduites à l’interface
Identification d'une loi de comportement plastique pour la Zone Affectée Thermiquement des soudures des aciers AHSS de 3ème génération
No full textNational audienceLe travail présenté propose une stratégie pour identifier la loi de comportement des Zones Affectées Thermiquement des soudures laser. Cette méthode s’appuie sur des essais de traction sur éprouvettes à encoches instrumentés avec un système de mesure de champs par corrélation d’images numériques adapté aux petites échelles. La loi de comportement dans la zone d’intérêt sera déterminée par approche inverse à partir des données expérimentales
Optimizing Microgrid Sizing, Energy Management, and Electric Vehicle Integration in Various French Cities
No full textInternational audienceEnsuring optimal sizing and efficient energy management in a microgrid is a complex optimization problem. This paper introduces a Mixed-Integer Linear Programming formulation designed to achieve a globally optimal solution for minimizing the microgrid Levelized Cost Of Energy (LCOE) while considering technical and economic factors. It also calculates life cycle emissions (LCE) and conducts a comparative analysis in various French cities. These microgrids incorporate PV panels, wind turbines, grid connections, and energy storage. The primary case study focuses on a university campus with electric vehicles. The results demonstrate that the combination of wind turbines and PV panels leads to a reduction in LCOE when compared to using either technology in isolation. Additionally, the research highlights that having higher average solar irradiation and wind speed does not always correlate with lower wind turbine and PV capacities, as the variability of data between seasons significantly influences the outcomes