HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
Co-optimisation du dimensionnement et du contrôle d'un micro-réseau urbain
No full textThe modernization of the electricity grid (EG) through the implementation of microgrids offers significant potential for enhancing energy resilience, sustainability, and efficiency. However, this transition involves navigating a complex web of technical, economic, and environmental challenges. Microgrids require meticulous planning and optimization to balance energygeneration, storage, and consumption while minimizing costs and carbon emissions. Achievingthis balance calls for advanced optimization strategies, which are capable of addressing theintricacies of system components and operational dynamics. The objective of this research is to enhance the decision-making capabilities of microgrid designers by providing a comprehensive approach for microgrid planning. The study offers an in-depth analysis of the project’s lifetime from technical, economic, and environmental perspectives. Implemented in Python and solved using CPLEX, the optimization process aims to minimize both the levelized cost of energy (LCOE) and the levelized cost of emissions (LCE). The study utilizes real economic and environmental data considering load growth as well as actual solar irradiation, ambient temperature, and wind speed data. The load for the university building is based on data from the Université de Technologie de Compiègne, France with the electric vehicle (EV) load modeled using probabilistic modeling. The study introduces a joint multi-objective optimization strategy usingMixed-Integer Linear Programming (MILP) to ensure globally optimal solutions, thereby that facilitates obtaining more informed and effective design choices. These choices involve evaluating various proposed solutions to balance cost and carbon emissions while addressing the complexities and technical constraints of the energy management (EM) problem. A novel aspect of this work is the integration of EM and component sizing into a unified optimization problem, aiming for an optimality gap of 0% with reduced computation time compared to existing literature. The proposed method evaluates the inherent trade-offs among various solutions by identifying the Pareto front and allowing for an optimal balance between economic and environmental objectives. The results indicate a significant reduction in LCOE and LCE in the GCMG compared to the IMG. The study reveals that Battery Energy Storage System (BESS) capacity increases as the LCE decreases, and the number of Photovoltaic (PV) systems is higher when the LCOE is lower for both operation modes. This occurs because the BESS has a slightly lower LCE compared to PV, and the LCOE of PV is also lower than that of BESS. Furthermore, as the limit of the EG increases, the Pareto fronts become lower and steeper. Additionally, the same MILP algorithm is applied to optimize microgrids from a tertiary university campus across various cities. The study further integrates wind turbines (WT) and EV loads into the microgrid. The study provides a comparative analysis of three scenarios (PV/BESS, WT/BESS, and PV/WT/BESS) across different cities for evaluating the impacts of seasonal fluctuations on LCOE and LCE, and for assessing how microgrid component technologies influence LCOE and LCE outcomes. The results indicate that scenarios including PV/WT/BESS yield the lowest LCOE and LCE values, while the WT/BESS scenario results in the highest LCOE and LCE. It is also observed that the order of cities based on average solar irradiation or wind speed does not necessarily correspond to the order of LCOE and LCE. Monthly and daily fluctuations in solar irradiation and wind speed significantly impact these results. Regarding the technologies, locally produced PV panels contribute positively to the overall LCE of the microgrid, with PV panels incorporating phase changing material showing higher LCE. The research also compares two distinct algorithms.La modernisation du réseau électrique (RE) par la mise en œuvre de micro-réseaux offre un potentiel significatif pour améliorer la résilience énergétique, la durabilité et l'efficacité. Cependant, cette transition implique de naviguer dans un ensemble complexe de défis techniques, économiques et environnementaux. Les micro-réseaux nécessitent une planification et une optimisation méticuleuses pour équilibrer la génération, le stockage et la consommation d'énergie tout en minimisant les coûts et les émissions de carbone. Atteindre cet équilibre nécessite des stratégies d'optimisation avancées capables de traiter les subtilités des composants du système et des dynamiques opérationnelles. L'objectif de cette recherche est d'améliorer les capacités de prise de décision des concepteurs de micro-réseaux en fournissant une approche globale pour la planification des micro-réseaux. L'étude offre une analyse approfondie du projet sur toute sa durée de vie, du point de vue technique, économique et environnemental. Implémenté en Python et résolu à l'aide de CPLEX, le processus d'optimisation vise à minimiser à la fois le « Levelized Cost of Energy » (LCOE) et le « Life Cycle Emission » (LCE). L'étude utilise des données économiques et environnementales réelles, en tenant compte de la croissance de la charge ainsi que des données réelles d'irradiation solaire, de température ambiante et de vitesse du vent. La charge pour le bâtiment universitaire est basée sur des données de l'Université de Technologie de Compiègne, France, avec la charge des véhicules électriques (VE) modélisée à l'aide d'une modélisation probabiliste. L'étude introduit une stratégie d'optimisation multi-objectifs conjointe utilisant le « Mixed Integer Linear Programming » (MILP) pour garantir des solutions globalement optimales, facilitant ainsi des choix de conception plus informés et efficaces. Ces choix impliquent l'évaluation de diverses solutions proposées pour équilibrer les coûts et les émissions de carbone tout en abordant les complexités et les contraintes techniques du problème de gestion de l'énergie. Un aspect novateur de ce travail est l'intégration de la gestion d'énergie (GE) et du dimensionnement des composants dans un problème d'optimisation unifié, visant un écart d'optimalité de 0 % avec un temps de calcul réduit par rapport à la littérature existante. La méthode proposée évalue les compromis inhérents entre diverses solutions, identifiant le front de Pareto et permettant un équilibre optimal entre les objectifs économiques et environnementaux. Les résultats indiquent une réduction significative du LCOE et du LCE dans le GCMG par rapport à l'IMG. L'étude révèle que la capacité du système de stockage d'énergie par batterie (BESS) augmente à mesure que le LCE diminue, et que le nombre de systèmes photovoltaïques est plus élevé lorsque le LCOE est plus bas pour les deux modes de fonctionnement. Cela se produit parce que le BESS a un LCE légèrement inférieur à celui du PV, et que le LCOE du PV est également inférieur à celui du BESS. De plus, à mesure que la limite du RE augmente, les fronts de Pareto deviennent plus bas et plus raides. En outre, le même algorithme MILP est appliqué pour optimiser les micro-réseaux d'un campus universitaire tertiaire dans diverses villes. L'étude intègre également des éoliennes (WT) et des charges de VE dans le micro-réseau. L'étude fournit une analyse comparative de trois scénarios (PV/BESS, WT/BESS et PV/WT/BESS) dans différentes villes, évalue les impacts des fluctuations saisonnières sur le LCOE et le LCE, et évalue comment les technologies des composants des micro-réseaux influencent les résultats du LCOE et du LCE. Les résultats indiquent que les scénarios incluant PV/WT/BESS produisent les valeurs de LCOE et de LCE les plus basses, tandis que le scénario WT/BESS aboutit aux valeurs de LCOE et de LCE les plus élevées
Single index regression for locally stationary functional time series
No full textInternational audienceIn this research, we formulated an asymptotic theory for single index regression applied to locally stationary functional time series. Our approach involved introducing estimators featuring a regression function that exhibited smooth temporal changes. We rigorously established the uniform convergence rates for kernel estimators, specifically the Nadaraya-Watson (NW) estimator for the regression function. Additionally, we provided a central limit theorem for the NW estimator. Finally, the theory was supported by a comprehensive simulation study to investigate the finite-sample performance of our proposed method.</p
Système de simulations de crises, basé sur une ontologie, pour la mise à l’abri des populations
No full textInternational audienceClimate change has led to an increase in the frequency and intensity of natural disasters, necessitating the development of efficient crisis management strategies for population sheltering. However, existing research on this topic primarily focuses on the use of public resources such as ambulances and fire trucks, which may sometimes be insufficient due to high demand and impacted locations, worsening the shortage of resources. This research introduces an ontology-based crisis simulation system for population sheltering management that focuses on the integration and distribution of citizen–volunteer drivers/vehicles into the evacuation process. Recognizing the limitations of public resources in current crisis management models, our approach incorporates citizen resources to enhance overall evacuation capacity. We develop an ontology to standardize crisis management knowledge, frame vehicle distribution as a recommendation problem, and design a simulation module incorporating a constraint-based recommender system. The proposed scenario illustrates how the simulation system can recommend citizen resources during crisis situations by considering the constraints to be satisfied. With our system, we aim at helping stakeholders to be prepared for various disaster scenarios: optimizing resource allocation and reducing time to make decisions by decision-makers
Guiding Urban Inclusivity: opening the public space in cities for women
No full textInternational audienceThis position paper identifies challenges that can be addressed through data-driven analytics to design inclusive urban spaces that meet women's security and well-being expectations. To this end, it proposes a decision-making support platform that enables urban planners to identify quantitative and qualitative variables that influence quality of life (QoL) and understand how much each variable contributes to perceptions of well-being
Automatic Image Annotation for Mapped Features Detection
No full textInternational audienceDetecting road features is a key enabler for autonomous driving and localization. For instance, a reliable detection of poles which are widespread in road environments can improve localization. Modern deep learning-based perception systems need a significant amount of annotated data. Automatic annotation avoids time-consuming and costly manual annotation. Because automatic methods are prone to errors, managing annotation uncertainty is crucial to ensure a proper learning process. Fusing multiple annotation sources on the same dataset can be an efficient way to reduce the errors. This not only improves the quality of annotations, but also improves the learning of perception models. In this paper, we consider the fusion of three automatic annotation methods in images: feature projection from a high accuracy vector map combined with a lidar, image segmentation and lidar segmentation. Our experimental results demonstrate the significant benefits of multi-modal automatic annotation for pole detection through a comparative evaluation on manually annotated images. Finally, the resulting multi-modal fusion is used to fine-tune an object detection model for pole base detection using unlabeled data, showing overall improvements achieved by enhancing network specialization. The dataset is publicly available
Réduction de modèles éléments finis volumiques par des approches volume-coque et volume-poutre ; application aux structures en bois multicouches
No full textNational audienceCet article présente des modèles éléments finis de poutres et de panneaux multicouches en bois assemblés à l'aide de goujons en bois densifié. À partir d'un modèle volumique considéré comme référence, une méthodologie utilisant une approche volume -poutre pour les goujons et volume -coque pour les lamelles est décrite, ce qui conduit à une réduction significative du nombre de degrés de liberté. Ces modèles sont évalués pour les performances vibratoires des structures assemblées en bois, et sont destinés à être utilisés pour une étude numérique de la variabilité
Hidden in Plain Sight: The Questions sur l’Encyclopédie in the Nineteenth Century
No full textInternational audienceThe question of Voltaire’s influence on the nineteenth century is famously thorny. The task is all the more complicated when considering a work like the Questions sur l’Encyclopédie, Voltaire’s longest and perhaps least-known masterpiece. For various editorial reasons, the Questions all but disappeared after their initial publication in the 1770s, a period during which they nonetheless had become a bestseller. Although no longer referred to using their rightful title, excerpts from the Questions proliferate in the print culture of postrevolutionary France, used and reused across an impressive swathe of texts of all stripes and sizes. The presence of these reuses would seem to corroborate the assertion that one important, if perhaps overlooked, aspect of Voltaire’s influence is his role as an agent of transmission. This role emphasises what might be called Voltaire’s aesthetic of reuse, and suggests a powerful new way of understanding this essential element of his participation in the Enlightenment project. This essay attempts to understand this particular version of Voltaire as an agent of transmission in the promulgation of Enlightenment ideas, while, at the same time, exploring the impact of his turn towards alphabetical works as a weapon of choice in the last decades of his life, as he engages in the larger arena of philosophical warfare opened up by the publication (and suppression) of the Encyclopédie in the 1750s. The embedding of questions so fundamental to the Enlightenment (religious toleration, philosophical and political liberty, human rights, fair treatment under the law, etc.) into Voltaire’s later dictionary works can perhaps help explain the immense success of the Questions, at least in the immediate years following their publication, and perhaps even in the postrevolutionary era, as these same questions will traverse the entire nineteenth century. With this in mind, our aim is to re-evaluate and reposition the Questions’ place (or absence) in literary history using current digital humanities methods and corpora to unearth references to Voltaire’s ‘hidden’ masterpiece of Enlightenment thought
Decision Making for Autonomous Vehicles based on Risk Assessment in a Dynamic Environment
No full textInternational audienceRecent advancements in autonomous driving technologies have significantly enhanced road safety and collision avoidance. However, ensuring this in dynamic driving environments remains a challenging endeavor. This paper addresses this challenge by proposing a high-level risk-aware decisionmaking module integrated into the trajectory planner of autonomous vehicles. The module establishes a function for dynamic risk assessment, considering both longitudinal and lateral aspects of the environment. By incorporating various factors such as velocity and relative position, the proposed function enables the vehicle to anticipate and respond to potential hazards proactively. Additionally, the paper incorporates and builds upon previous work of a modular and distinctive AIbased Adaptive Cruise Control (ACC) system with robust generalization capabilities. Results demonstrate the effectiveness of the proposed approach in improving safety, collision avoidance, and adaptability in highly interactive driving environments allowing the vehicle to dynamically re-plan trajectories and speed profiles during lane change maneuvers.</div
Safe and Energy-Efficient Jerk-Controlled Speed Profiling for On-Road Autonomous Vehicles
No full textInternational audienceEfficient speed planning is crucial for the safe and comfortable navigation of autonomous vehicles in dynamic environments. This paper introduces a novel energy-efficient, jerk-controlled speed planning approach based on quintic polynomial generation. We present a systematic methodology to determine the dynamic speed of autonomous vehicles by integrating several factors, including the relative velocity with dynamic obstacles, the curvature of the base frame and optimal selected path, road adherence, and road gradient. The direct integration of road adherence and gradient into the speed profiling approach contributes to improving vehicle safety. Comparative analysis with literature methods demonstrates the significant impact of jerk smoothness on energy efficiency. Simulations are conducted in a joint simulation between Simulink/Matlab and SCANeR Studio vehicle dynamics simulator, followed by validation on a real-world dataset. Our findings elucidate the significance of the proposed planning method in enhancing safety, energy economy, driving comfort, and computational efficiency, while effectively addressing a wide range of critical situations
Density analysis based on normalized process diagrams : a comparative review of ferromagnetic alloys produced by laser powder bed fusion
No full textInternational audienceThis research explores additive manufacturing of magnetic materials via Selective Laser Melting (SLM), integrating insights from prior studies. Emphasizing density variation concerning applied energies, our methodology incorporates normalized process diagrams. This analysis revealing an identification of different density regions in the case of Fe-6.5%wt. Si alloy. These regions show the interplay between process parameters and thermophysical material properties, encompassing lack of fusion, keyhole pores and balling. Further, density variations of Fe-3% wt. Si, Fe-6.7%wt. Si, and Fe-6.9%wt. Si alloys are studied, highlighting the impact of silicon content on final density. The present study contributes to understand dynamic interactions shaping magnetic material density. By optimizing process parameters based on density considerations, microstructure tailoring becomes possible, and, so far, an optimization of the magnetic properties