Université de technologie de Troyes open archive
Not a member yet
10722 research outputs found
Sort by
Brownian-Inspired Neural Networks for Explainable Edge Vibration Anomaly Detection
No full textEmbedded vibration sensors enable real-time anomaly detection in remote and resource-constrained environments, yet existing methods often lack adaptability and interpretability. We propose a physics-informed Brownian-inspired neural network that models cumulative vibration features as stochastic trajectories and introduces Stochastic Brownian activation (SABA). Unlike standard activations, SABA embeds Brownian statistics to yield bounded, adaptive, and explainable responses directly linked to mechanical principles of vibrations. A lightweight architecture demonstrates effective preliminary results on real data collected from sensitive infrastructures, providing calibrated anomaly scores suitable for embedded deployment. This work opens a new direction by combining stochastic physics with efficient edge AI for interpretable vibration anomaly detection
Structure and conducting properties of poly(3-butylthiophene) nanometer-thick films:influence of the doping method
No full textInternational audienc
Bright-Field Polarimetry of a Single Plasmonic Nanostructure Combining Polarization and Position Modulation Techniques
No full textInternational audienceBeing able to measure the chiroptical properties or more generally the specific response of individual nanoparticles with respect to the polarization of light is a major challenge in the domain of nanophotonics, whether for a fundamental purpose or for understanding and shaping the properties of metamaterials built from these entities. The only few published experiments in this field are essentially dedicated to identifying the signature of circular dichroism. On the basis of conventional methods employed in the polarimetric study of macroscopic samples, we propose an alternative experimental technique which allows a complete determination of the relevant optical anisotropy parameters of a single nano-object (circular and linear dichroisms and birefringences). The retrieval of weak signals from bright-field extinction measurements is made possible by the original combination of polarization and spatial modulation spectroscopies. The framework and applicability of the method and data processing are discussed in detail. They are illustrated through the study of two different types of nanoparticles: achiral homodimers of gold nanospheres and chiral gold nanostructures lithographied on a transparent substrate
Pathway to degrowth in design practices: the tensions between engineer and tinkerer logics in low-tech design projects
No full textInternational audienceFor the past ten years or so, the Low-Tech movement has been proposing a rethink of the way technologies are designed, and promoting a paradigm shift towards degrowth or post-growth societies through design. However, there are many different visions within this movement, leading to different ways of practising Low-Tech design. On the one hand, some actors are closer to the design practices of the growth paradigm, mainly following the logic of the engineer*, focused on the search for performance, particularly economic and technical performance. On the other hand, some actors move further away from this paradigm, mainly following a bricoleur (tinkerer)* logic, based on ingenuity and the ability to find solutions with few or no resources. The aim of this qualitative research is to gain a better understanding of these differences and to assess their effectiveness in terms of sustainability of the solutions obtained.The data was collected during 5 participant observations on Low-Tech design projects between October 2024 and March 2025. The observations took place at different stages of the design process, depending on the project (conceptualisation, prototyping, testing, etc.) and lasted between 3 and 5 days. Additional semi-structured interviews lasting around 60 minutes were conducted with the designers to gather data from the design phases prior to the observation phase. To identify the use of engineer and bricoleur logics, a series of markers for the use of these logics were defined on the basis of the literature on the subject. The design thinking framework was also used to compare the use of these logics throughout the duration of the projects. Finally, the sustainable potential of the various solutions resulting from the projects observed was assessed and compared using the Convivial Technologies Matrix (Vetter 2017). The aim of this work is to show the different ways in which the logic of the engineer and the logic of the bricoleur can interact within a low-tech project, and to assess whether some of these ways of working produce more sustainable solutions than others. Ultimately, these results will help to show the possibilities, within a movement promoting a paradigm shift, of striking a balance between practices that are vestiges of the current paradigm and singular experiments projecting towards new horizons.*Concepts developed by:Levi-Strauss, Claude. 1962. « La Pensée sauvage ». In , Plon, 26‑33.And recently brought up to date by:Lederlin, Fanny. 2023. Éloge du bricolage : soucis des choses, soin des vivants et liberté d’agir. Puf.Vetter, A., 2017. The Matrix of Convivial Technology–Assessing technologies for degrowth. Journal of Cleaner Production.Depuis une dizaine d’années, le mouvement Low-Tech propose de repenser la manière de concevoir les technologies et promeut un changement de paradigme vers des sociétés en décroissance ou post-croissance grâce à la conception. Cependant, il existe une pluralité de visions au sein de ce mouvement induisant différentes manières de pratiquer la conception Low-Tech. En effet, d’un côté certains acteurs se veulent plus proche des pratiques de conception du paradigme de la croissance en suivant majoritairement une logique de l’ingénieur*, axée sur la recherche de performance, notamment économique et technique. D’un autre côté, certains acteurs s’en éloignent plus radicalement en suivant majoritairement une logique du bricoleur*, axée sur l’ingéniosité, la capacité à trouver des solutions avec peu ou pas de ressources. Ce travail de recherche qualitatif propose de mieux comprendre ces différences de fonctionnement et d’évaluer leur efficacité en terme de potentiel de soutenabilité des solutions obtenues.Les données ont été récoltées lors de 5 observations participantes sur des projets de conception Low-Tech entre octobre 2024 et mars 2025. Les observations sont intervenues à différentes phases du processus de conception selon les projets (conceptualisation, prototypage, tests…) et leurs durées varient entre 3 et 5 jours. Des entretiens semi-directifs complémentaires d’environ 60 minutes ont été réalisés avec les concepteurs pour récolter les données des phases de conception amont de celle de l’observation. Pour identifier l’utilisation des logiques de l’ingénieur et du bricoleur, une série de marqueur d’utilisation de ces logiques ont été définis à partir de la littérature sur le sujet. La cadre du design thinking a également été utilisé pour comparer l’utilisation de ces logiques sur toute la temporalité des projets. Enfin, le potentiel soutenable des différentes solutions résultant des projets observés a été évalué et comparé en utilisation la matrice des technologies conviviales (Vetter 2017). L’objectif de ce travail est de montrer les différentes possibilités d’interactions entre logique de l’ingénieur et logique du bricoleur au sein d’un projet Low-Tech et d’évaluer si certains de ces fonctionnements produisent des solutions plus soutenables que d’autres. A terme, ces résultats contribueront à montrer les possibles, au sein d’un mouvement promouvant un changement de paradigme, d’équilibre entre pratiques vestiges du paradigme actuel et expérimentations singulières projetant vers un nouvel horizon.*Concepts développés par : Levi-Strauss, Claude. 1962. « La Pensée sauvage ». In , Plon, 26‑33.Et repris récemment par : Lederlin, Fanny. 2023. Éloge du bricolage : soucis des choses, soin des vivants et liberté d’agir. Puf.Vetter, A., 2017. The Matrix of Convivial Technology–Assessing technologies for degrowth. Journal of Cleaner Production
Méthodes approchées pour le partage de livraisons dans les circuits courts alimentaires
No full textInternational audienc
Les perturbations de la coordination au sein d’une organisation temporaire particulière : les cellules de crise
No full textInternational audienc
The joint stochastic multi-period lot sizing and two-dimensional variable-sized cutting stock problem
No full textInternational audienc
Failure Causality Diagnostic in Industrial Systems through Automated Machine Learning
No full textInternational audienceIn the context of modern industrial systems, efficient failure management is crucial for maintaining operational integrity, minimizing downtime, and maintenance optimization. This paper explores the application of Automated Machine Learning (AutoML) to enhance both failure causality diagnostic and failure causality prognostic in industrial systems. Different failure causes are detected by failure causality diagnostics, and the upcoming failure could be prevented by failure causality prognostics. Indeed, upcoming failures could be avoided by preventing their causalities.Traditional machine learning approaches require significant manual intervention for model selection, hyperparameter tuning, and feature engineering, which can be time and cost-consuming. AutoML, on the other hand, automates these processes, enabling quicker and more accurate predictions while reducing the need for extensive domain expertise.Our approach integrates AutoML into real-time failure diagnostics, identifying the root causes of system malfunctions using historical and sensor data. Simultaneously, it applies AutoML for prognostics, predicting Remaining Useful Life (RUL) of components and foreseeing future failures. By leveraging both data-driven models and physics-based insights, our approach improves the reliability of diagnostics and prognostics in various industries, including manufacturing, aerospace, and energy.The competing risks can be considered an application of this approach. High probable competing risks to cause are detected based on historical data in diagnostic phase and by handling them in the prognostic phase, the anticipated failures could be prevented.Finally, the experiments are conducted using real-world industrial datasets, demonstrating the superior performance of AutoML compared to traditional machine learning methods in both diagnostic accuracy and prognostic precision. This study shows that AutoML can significantly enhance decision-making processes in maintenance planning and risk mitigation, ultimately reducing operational costs and improving system reliability
Simple Detection of AI-Generated Images based on Noise Correlation
No full textInternational audienceThe present paper deals with the problem of the detection of AI-generated images. It first proposes a forensic analysis, based on spatial correlations of the noise present in images, that can be used as fingerprints of both real and generated images. In particular, fingerprints can be extracted in each color channel, and complement each other during detection. The proposed detection scheme is a 3-step classifier, consisting only of a set of simple log-linear classifiers. This scheme is shown to perform much better than a standalone detector. The performance of the method is first assessed in an In-Distribution scenario, where an error probability of less than 1% is achieved on uncompressed images. It is then compared to stateof-the-art detectors in an out-of-distribution scenario, where significant performance gains are achieved. Results highlight the good generalization performances to unseen generators and the liability of color channels, specifically the chrominance CbCr for current state-of-the-art generators. A robustness analysis to JPEG compression also shows promising results for our method.</div
Mathematical modeling of solar farm performance degradation in a dynamic environment for condition-based maintenance
No full textInternational audienceThis paper aims to address the challenge of modeling and optimizing condition-based maintenance policies for a degraded solar farm in varying environmental conditions. Dust accumulation and temperature increases are the two main causes of performance reduction and energy loss in the system. In this research, dust accumulation is modeled by the non-homogeneous compound Poisson process, and three different mathematical models for the efficiency reduction of photovoltaic panels due to dust accumulation are considered. The effects of wind and rain, taken as covariates, on dust accumulation and temperature are investigated by stochastic process modeling. The covariate process is considered a time-homogeneous Markov chain with finite state space. The PV surface temperature is modeled by a non-homogeneous Markov chain with finite state space and transition matrices under covariate states. Different PV panels exhibit varied degradation rates, influenced by their position and tilt angle to sunlight. In the framework of the system, we derive multiple maintenance policies aimed at achieving the minimum cost criterion. The expected long-term average maintenance costs under different covariate conditions and maintenance policies are evaluated through simulation experiments to compare the effectiveness of each policy