HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
Optimisation des procédés solides divisés par la maîtrise des propriétés d’écoulement et des phénomènes de transfert fluide-solide
No full textMéthodes et outils pour soutenir l’éco-conception dans un laboratoire d’innovation industriel
No full textTo remain competitive, companies must be innovative and regularly offer new products, services, processes, etc. If, for a long-time, innovation has focused on the technical dimension, today it relies on three main levers: prices, technology and uses. Recently, sustainable innovation, which consists in innovating while considering the environmental impacts, is also considered as an additional innovation lever. To meet these new demands, particularly around innovation through use and sustainable innovation, companies are no longer wondering if they should innovate but rather how to do so. They “seek to renew their methods of innovation and creation”. New methods and tools are deployed in an industrial context. One of the practices proposed is the establishment of new places, dedicated to innovation, inspired by the maker movement such as FabLabs, Living labs, Hackerspaces, Makerspaces, ... These new spaces, supported by various actors, renew the modalities of innovation and creation through the implementation of collaborative and iterative processes. Their objectives are to put uses back at the heart of innovation processes, to focus the process on the user and his needs, to give a new impetus to the exploration and innovation processes of companies, to upgrade practical skills and adapt to a context of deindustrialization. To achieve these objectives, these innovation laboratories are based on three main pillars: a place characterized by a particular architecture, layout and decor that influence the behavior of participants; a team composed of diverse and heterogeneous actors such as researchers, engineers as well as experts in methods and tools of creativity and prototyping and finally methods to facilitate and support the generation of ideas and group work, notably through co-creativity sessions. In the context of sustainable innovation, very few tools are adapted to the practices of these new places of innovation. Indeed, innovation laboratories support the generation of ideas and first concepts while eco-innovation or eco-design tools such as Life Cycle Assessment, are based on the assessment of the environmental impacts of products. We propose to develop an eco-creativity tool that aims to accompany the participants of a co-creativity session to generate new ideas, original, respectful of environment and adapted to the context in which they manifest. To evaluate the effects of the introduction of such a tool during co-creativity sessions, a comprehensive analysis methodology of these sessions was developed. To carry out this work, we have integrated the Clean Mobility LAB, innovation laboratory of a large international automotive equipment engineering and production group, FORVIA FAURECIAAfin de rester compétitives, les entreprises se doivent d’être innovantes et de proposer régulièrement de nouveaux produits, services, process, … Si, pendant longtemps, les innovations se sont concentrées sur la dimension technique, aujourd’hui, elles s’appuient sur trois leviers principaux : les prix, la technologie et les usages. Depuis peu, l’innovation durable, qui consiste à innover en prenant en compte les impacts environnementaux, est également considérée comme un levier d’innovation supplémentaire. Pour répondre à ces nouvelles demandes, notamment autour de l’innovation par l’usage et l’innovation durable, les entreprises ne se demandent plus si elles doivent innover mais plutôt comment le faire. Elles « cherchent à renouveler leurs modalités d’innovation et de création ». De nouvelles méthodes et de nouveaux outils sont déployés en contexte industriel. Une des pratiques proposées est la mise en place de nouveaux lieux, dédiés à l’innovation, inspirés du mouvement maker tels que les FabLabs, les Living labs, les Hackerspaces, les Makerspaces, ... Ces nouveaux espaces, portés par des acteurs divers, renouvellent les modalités d’innovation et de création par la mise en oeuvre de processus collaboratifs et itératifs. Ils ont pour objectifs de remettre les usages au coeur des processus d’innovation, de centrer le processus sur l’utilisateur et ses besoins, de donner un nouveau souffle aux processus d’exploration et d’innovation des entreprises, de revaloriser les compétences pratiques et de s’adapter à un contexte de désindustrialisation. Pour réaliser ces objectifs, ces laboratoires d’innovation s’appuient sur trois grands piliers : un lieu caractérisé par une architecture, une disposition et un décor particuliers qui influencent le comportement des occupants ; une équipe composée d’acteurs variés et hétérogènes comme des chercheurs, des ingénieurs ainsi que des experts en méthodes et outils de créativité et de prototypage et enfin des méthodes visant à faciliter et accompagner la génération d’idées et le travail de groupe, notamment à travers des sessions de co-créativité. Dans le cadre de l’innovation durable, très peu d’outils sont adaptés aux pratiques de ces nouveaux lieux d’innovation. En effet, les laboratoires d’innovation soutiennent la génération d’idées et de premiers concepts alors que les outils d’éco-innovation ou d’éco-conception tel que l’Analyse de Cycle de Vie, reposent sur l’évaluation des impacts environnementaux de produits finalisés. Nous proposons de développer un outil d’éco-créativité qui a pour objectif d’accompagner les participants d’une session de co-créativité pour générer des idées nouvelles, originales, respectueuses de l’environnement et adaptées au contexte dans lequel elles se manifestent. Pour évaluer les effets de l’introduction d’un tel outil pendant des séances de co-créativité, une méthodologie d’analyse exhaustive de ces séances a été développée. Pour réaliser ces travaux, nous avons intégré le Clean Mobility LAB, laboratoire d’innovation d’un grand groupe d’ingénierie et de production d’équipements automobiles international, FORVIA FAURECIA
Investigation of Oleic Acid, Palmitic Acid and their mixture on the development of hepatic steatosis using liver-on-chip technology
No full textInternational audienc
Quantified analysis of facial movement: a reference for clinical applications
No full textInternational audienceMost techniques for evaluating unilateral impairments in facial movement yield subjective measurements. The objective of the present study was to define a reference dataset and develop a visualisation tool for clinical assessments. In this prospective study, a motion capture system was used to quantify facial movements in 30 healthy adults and 2 patients. We analysed the displacements of 105 reflective markers placed on the participant's face during five movements (M1-M5). For each marker, the primary endpoint was the maximum amplitude of displacement from the static position (M0) in an analysis of variance. The measurement precision was 0.1 mm. Significant displacements of markers were identified for M1-M5, and displacement patterns were defined. The patients and age-matched healthy participants were compared with regard to the amplitude of displacement. We created a new type of radar plot to visually represent the diagnosis and facilitate effective communication between medical professionals. In proof-of-concept experiments, we collected quantitative data on patients with facial palsy and created a patient-specific radar plot. Our new protocol for clinical facial motion capture ("quantified analysis of facial movement", QAFM) was accurate and should thus facilitate the long-term clinical follow-up of patients with facial palsy. To take account of the limitations affecting the comparison with the healthy side, we created a dataset of healthy facial movements; our method might therefore be applicable to other conditions in which movements on one or both sides of the face are impaired. The patient-specific radar plot enables clinicians to read and understand the results rapidly
Damping loss factor characterization of complex structures using a Green’s function-based model
No full textInternational audienceThe Inhomogeneous Wave Correlation technique has been used to estimate experimentally the wavenumber of simple and complex structures. However this method is essentially focused on the real part of the wavenumber and is not always robust to measure the damping loss factor related to its imaginary part. Another correlation technique able to capture the damping effects of complex structures as function of the heading angle accurately is presented here. Instead of using an inhomogeneous plane wave, the correlation is made with a Green’s function to better take into account the behavior of the structure excited by a point force. Several results, with simulated and measured data, are compared with an analytical discrete laminate model and show the accuracy of this Green’s Function Correlation technique to recover the damping loss factor of complex structures with high damping and in the mid to high frequency range
Ustekinumab Use in Pediatric Inflammatory Bowel Disease: A French Multicenter Study From the Pediatric GETAID
No full textInternational audienceObjectives: Ustekinumab is known to be efficient in adult patients suffering from moderate to severe Crohn disease (CD) and ulcerative colitis (UC) resistant to anti-tumor necrosis factor-alpha (TNF-α). Here, we described the clinical course of treatment with ustekinumab in French pediatric inflammatory bowel disease (IBD) patients treated with ustekinumab. Methods: This study includes all pediatric patients treated by ustekinumab injection for IBD (CD and UC), between January 2016 and December 2019. Results: Fifty-three patients were enrolled, 15 males and 38 females. Forty-eight patients (90%) had a diagnosis of CD and 5 (9.4%) had UC. Sixty-five percent of CD patients presented an ileocolitis. Perineal disease was observed in 20 out of 48 CD patients (41.7%), among them 9 were treated surgically. All patients included were resistant to anti-TNF-α treatment. Fifty-one percent had presented side effects linked to anti-TNF-α, including psoriasis and anaphylactic reaction. The average Pediatric Crohn Disease Activity Index (PCDAI) at induction was 28.7 (5–85), 18.7 (0–75) at 3 months of treatment and 10 (0–35) at the last follow-up. The average Pediatric Ulcerative Colitis Activity Index at induction was 47 (25–65), 25 (15–40) at 3 months of treatment and 18.3 (0–35) at the last follow-up. No severe side effects were observed. Conclusion: In this retrospective, multicentral study, ustekinumab proved to be efficient in pediatric patients resistant to anti-TNF-α. PCDAI has been significantly improved in patients with severe disease, treated with ustekinumab
Recommandation de ressources au sein de plateformes e-learning basée sur un graphe de connaissances
No full textAs an integral part of education, e-learning provides users with massive resources and allows them learn independently without the constraints of time and space. However, flooded with such a huge number of resources offered by different individuals and institutions, personalized learning is required to help users choose resources precisely. Recommender systems are widely used in e-learning platforms to reduce the information burden and achieve personalized learning. Generally speaking, the more information a recommender system knows, the more accurate the recommendations will be. But conventional recommendation approaches usually ignore a pivotal piece of information, the latent connections between entities in the e-learning platform. Collaborative filtering approaches make recommendations by learning user-resource historical interactions (e.g., user ratings and browsing history), content-based approaches mainly consider the relevant information of users and resources (e.g., user profile and resource content). In addition to user and resource, there are numerousentities lurking in e-learning platform, and the latent connections between these entities can provide crucial information support to the function of recommender system. For example, users may choose resources created by authors with whom they are familiar, i.e., there are latent connections between users and authors, and the connections affect users’ choices. Graph based technologies (e.g., knowledge graph) have the capacity to extract, represent, manipulate, and model information (including latent connections) in different domains. In this context, we propose a recommendation approach incorporating knowledge graph to recommend pedagogical resources to e-learning platform users. We integrate knowledge graph, feature extraction and neural network into a recommendation framework. It contains three modules: i) modeling the information of e-learning platform via knowledge graph; ii) acquiring features of users and resources from the formed knowledge graph; iii) learning the acquired features with neural network for resource recommendation. To evaluate the performance of the recommendation framework, we conduct a series of experiments based on two datasets from real-world e-learning platforms, and the results confirm that the proposed recommendation framework outperforms the methods from the literature.En tant que partie intégrante de l’éducation, l’e-learning fournit aux utilisateurs des ressources considérables et leur permet d’apprendre de manière autonome sans les contraintes de temps et d’espace. Cependant, face au grand nombre de ressources proposées par différents individus et institutions, une aide personnalisée au choix de ces dernières semble nécessaire pour faciliter l’apprentissage. Les systèmes de recommandation sont largement utilisés dans les plateformes e-learning pour réduire la surcharge d’information et parvenir à l’apprentissage personnalisé. En général, plus un système d’information connait d’informations plus les recommandations seront pertinentes. Mais les approches de recommandation conventionnelles ignorent généralement les informations issues des relations que peuvent entretenir les entités utilisées par les plateformes e-learning. Les approches de filtrage collaboratif font des recommandations en apprenant les interactions historiques utilisateur-ressource (par exemple, les évaluations des utilisateurs et l’historique de navigation), les approches basées sur le contenu prennent principalement en compte les informations pertinentes sur les utilisateurs et les ressources (par exemple, le profil utilisateur et le contenu des ressources). En plus de l’utilisateur et de la ressource, de nombreuses entités se cachent dans la plateforme e-learning, et les relations que peuvent entretenir ces entités et le support informationnel qu’elles peuvent offrir à un système de recommandation. Par exemple, un utilisateur peut préférer choisir des ressources créées par un auteur spécifique, c’est-à-dire qu’il existe une connexion latente entre l’utilisateur et l’auteur, et cette connexion affecte les choix de l’utilisateur. Les technologies basées sur les graphes (par exemple, le graphe de connaissance) ont la capacité d’extraire, de représenter, de manipuler et de modéliser des informations dans différents domaines. Dans ce contexte, nous proposons une approche de recommandation intégrant un graphe de connaissance pour recommander des ressources pédagogiques aux utilisateurs de plateformes e-learning. Nous intégrons le graphe de connaissance, l’extraction de fonctionnalités et le réseau de neurones dans un cadre de recommandation. Un tel graphe comporte trois modules : i) modélisation des informations de la plateforme e-learning via un graphe de connaissances; ii) acquérir des fonctionnalités d’utilisateurs et de ressources à partir du graphe de connaissances formé; iii) apprendre les fonctionnalités acquises avec réseau de neurones pour la recommandation de ressources. Pour évaluer les performances du cadre de recommandation, nous conduisons une série d’expériences basées sur deux ensembles de données provenant de plateformes e-learning du monde réel, et les résultats confirment que le cadre de recommandation proposé surpasse les méthodes de la littérature
Eco-friendly green approach of nickel oxide nanoparticles for biomedical applications
No full textInternational audienceThe two most prominent areas of nanotechnology and nanoscience are environmental remediation and biomedical applications, which has a strong need to develop cleaner and more eco-friendly methods for preparing functional nanomaterials. This involves incorporating non-toxic chemicals and reagents for converting metal cations into metal or metal oxide nanoparticles (NPs), using readily available plant reagents and microbes as redox mediators. The extracts of these plants and microbes contain varying amounts of phytochemicals and metabolites that act as redox mediators and capping agents to stabilize biosynthesized NPs. Considering these natural reagents for forming metal/metal oxide NPs, the present work aims to provide a complete review of the green fabrication of nickel oxide (NiO) NPs using extracts from various plant and microbial sources. In addition, the performance of various biosynthesized NiO NPs and their potential uses in biological applications are discussed
SOME TRENDS OF OPERATIONS RESEARCH WITHA SPECIAL FOCUS ON ROBUST OPTIMISATION
No full textInternational audienc
Mesoscopic segregation in H13 steel molten pool during laser remelting: A combined influence of Marangoni convection and oxidation
No full textInternational audienceLaser-assisted metal processing has received great attention for repair, manufacturing, and surface modification in various industries. However, the mesoscopic segregation and oxidation phenomena during these processes require further investigations. In present work, the H13 tool steel is remelted by a laser beam. The generation of oxide layer, the profile and the mesoscopic segregation of the molten pool, Marangoni convection, the and hardening effect are studied by an experimentation combined with a multiphysics coupled simulation. An oxide layer was formed on the remelted surface with a thickness of ∼ 60 µm in the middle and ∼10 µm in the edges. The lightweight elements (Si, V, Cr, and Mn) float upon the molten pool driven by both buoyancy and Marangoni convection. In addition, due to their higher redox activity, they tend to have a higher proportion in the oxide layer with respect to the base metal. As a consequence, elemental mesoscopic segregation appeared in the remelted zone. The oxide layer shows an average hardness of 11.8 GPa, which is similar to that of the remelted zone (11.7 GPa). However, the brittleness of the oxide layer leads to cracks, hence deteriorating its mechanical properties. Together, these results bring a further fundamental understanding of the mesoscopic segregation that is attributed to the Marangoni convection and oxidation phenomenon during the laser processing of alloys, which might be a critical factor to optimize the mechanical properties