Portail HAL edf
Not a member yet
11029 research outputs found
Sort by
Toward Brague river flood modelling 3: the impact of culvert representation
No full textInternational audienceThe Brague River (south-east France) has historically been subjected to significant flooding events, notably recently in 2011, 2015 and 2019. This research primarily focuses on the 2015 flood event, which resulted in human loss and significant damage. In the study domain, several hydraulic structures across the river (e.g., under the A8 highway), which can hinder the river flow during extreme flood events. Our objective in this work is to assess the influence of culvert representation within the domain on hydraulic discharge, particularly beneath roadways, and to examine the resultant variations in flood extents and water depth. To this goal, the TELEMAC-2D hydraulic modelling software is used to simulate a range of modelling approaches that account differently for the flow in the culvert areas, such as opened or closed topography or using culvert discharge equations. Five scenarios concerning the highway culverts are proposed. The numerical results are analysed and intercompared with field observations. The results reveal a significant reduction in culverts flow rate upon the formation of logjams, while the highway overflow increases. These studies' findings could help in guiding management policies and practices, offering a scientific foundation for decision-making in the context of flood events involving hydraulic structures
Prévision de séries temporelles sous contraintes
No full textTime series forecasting presents unique challenges that limit the effectiveness of traditional machine learning algorithms. To address these limitations, various approaches have incorporated linear constraints into learning algorithms, such as generalized additive models and hierarchical forecasting. In this paper, we propose a unified framework for integrating and combining linear constraints in time series forecasting. Within this framework, we show that the exact minimizer of the constrained empirical risk can be computed efficiently using linear algebra alone. This approach allows for highly scalable implementations optimized for GPUs. We validate the proposed methodology through extensive benchmarking on real-world tasks, including electricity demand forecasting and tourism forecasting, achieving state-of-the-art performance.La prévision des séries temporelles présente des défis uniques qui limitent l'efficacité des algorithmes traditionnels d'apprentissage automatique. Pour y remédier, diverses approches cherchent à intégrer des contraintes linéaires dans les algorithmes d'apprentissage, tels que les modèles additifs généralisés et les prévisions hiérarchiques. Dans cet article, nous proposons un cadre unifié pour l'intégration et la combinaison des contraintes linéaires dans la prévision des séries temporelles. Dans ce cadre, nous montrons que le minimiseur exact du risque empirique contraint peut être calculé efficacement en utilisant uniquement l'algèbre linéaire. Cette approche permet l'implémentation efficace d'un algorithme sur carte graphique (GPU). Nous validons la méthodologie proposée par des expériences sur données réelles, en prévision de demande électrique et de tourisme
Development of an offline and online hybrid model for the Integrated Forecasting System
No full textInternational audienceIn recent years, there has been significant progress in the development of fully data‐driven global numerical weather prediction models. These machine learning weather prediction models have their strengths, notably accuracy and low computational requirements, but also their weaknesses: they struggle to represent fundamental dynamical balances, and they are far from being suitable for data assimilation experiments. Hybrid modelling emerges as a promising approach to address these limitations. Hybrid models integrate a physics‐based core component with a statistical component, typically a neural network, to enhance prediction capabilities. In this article, we propose to develop a model‐error correction for the operational Integrated Forecasting System (IFS) of the European Centre for Medium‐Range Weather Forecasts using a neural network. The neural network is initially pre‐trained offline using a large dataset of operational analyses and analysis increments. Subsequently, the trained network is integrated into the IFS within the Object‐Oriented Prediction System (OOPS) so as to be used in data assimilation and forecast experiments. It is then trained further online using a recently developed variant of weak‐constraint 4D‐Var. The results show that the pre‐trained neural network already provides a reliable model‐error correction, which translates into reduced forecast errors in many conditions, and that the online training improves the accuracy of the hybrid model further in many conditions
Local-scale experimental investigation of a two-phase cross-flow in a tube bundle and flow-induced vibration: Intermittent/Churn flow
No full textInternational audienceTwo-phase cross-flows in tube bundle can induce vibrations. This occurs in many industrial situations, such as nuclear power plant U-tube Steam Generator (SG). With the purpose of gathering high quality data for the validation of multiphase CFD simulation tools, a new experimental apparatus was designed and put in operation. The facility is instrumented for the two-phase air-water flow and tube vibration characterization. This article deals with experimental results in the churn/intermittent two-phase flow conditions. The experimental study was achieved by means of three measurement techniques: high-speed camera, optical dual-tip probe and wire mesh sensor. The behavior of the two-phase flow upstream of the tube bundle, characterized by the continuous generation of large gas structures, is detailed: highspeed camera and wire mesh sensor data reveal the prevalence of dispersed bubbles near the wall and large gas structures in the core region. The rising frequency of the large gas structures could be determined. Profiles of void fraction and gas velocity were measured within the tube bundle.The flow regime map of the experiment is proposed and is correlated with the study of flow-induced tube vibration. It highlights the role played by gas structure size distribution, thus flow pattern in the structural response in intermittent flow regimes with a local-scale characterization of the flow. Finally, the present work is a major contribution to improve and validate local-scale numerical simulation
Interfacial reactions between atomic layer deposited NiO x hole transport layers and metal halide perovskites in n-i-p perovskite solar cells
No full textInternational audienceWe observe defect formation at the metal halide perovskite (MHP)/ALD-NiO x interface, as revealed by HAXPES. Introduction of an organic buffer layer minimizes these defects, leading to improved device performance
Station automatique de jaugeage par dilution au selRetour d’expérience et bilan des installations en France et dans le monde
No full textInternational audienceThe article presents the use of AutoSalt (Fathom Scientific), an automatic salt dilution measurement device, in various contexts in France by different actors involved in hydrometric measurement. It aplpies the salt dilution gauging method with a conductivitymeter, which is a well-established method for measuring flow in well-mixed watercourses ([1] Dumas 1952, [3] Moore 2004, [4] Richardson 2017). The specific features of Autosalt, designed to automatically perform gauging on targeted flow ranges, are presented, followed by a wide range of sites that have been instrumented in recent years.L'article présente l’utilisation de l’AutoSalt (Fathom Scientific), qui est un dispositif de mesure automatique par dilution au sel dans différents contextes d’utilisation en France par différents acteurs de la mesure en hydrométrie. Il fonctionne par application de la méthode de jaugeage par dilution au sel avec un conductimètre selon la méthode dite « globale » ou « instantanée » qui est une méthode bien établie pour mesurer le débit dans les cours d'eaubien brassés ([1] Dumas 1952, [3] Moore 2004, [4] Richardson 2017). Sont respectivement présentés les spécificités de l’Autosalt, conçu pour réaliser automatiquement des jaugeages sur des plages de débit ciblées, puis un large panel de sites instrumentés ces dernières années
Following the equilibrium path during crack propagation under quasi-static loads: linear elastic fracture mechanics and phase-field approaches.
No full textInternational audienceThe equilibrium path [1] of an elastic body with an initial crack represents the sequence of states (displacement and crack geometry) at which the crack propagation threshold is met under a specific quasi-static proportional loading sequence. Constructing this equilibrium path requires adjusting the load amplitude to maintain the crack at its propagation threshold throughout its progression. Following the equilibrium path yields stable, incremental crack propagation at the cost of neglecting dynamic effects arising during instabilities. Following the equilibrium path in numerical simulations (e.g., phase-field fracture models) ensures controlled, incremental crack propagation, with each crack increment minimizing energy (in contrast to minimizing energy over the entire crack path).This work proposes methodologies for determining the equilibrium path using the frameworks of Linear Elastic Fracture Mechanics (LEFM) and phase-field fracture models. It focuses on a single controlled load, but extensions to more complex loads will also be discussed.We first examine the analytical case of a finite elastic body with a known crack path. This simple example introduces the core principles of path-following methods in fracture mechanics, serving as a foundation for more complex scenarios. We then address cases where the crack path is unknown a priori, requiring adaptations of the methodology to integrate numerical tools for elastic solutions, Stress Intensity Factor (SIF) computation, and crack-propagation solutions. Finally, we present the determination of the equilibrium path using phase-field fracture models, employing an extension [2] of arc-length methods.REFERENCES [1] Riks, E. (1979). An incremental approach to the solution of snapping and buckling problems. International Journal of Solids and Structures, 15(7), 529–551. [2] Chen, Z., & Schreyer, H. L. (1991). Secant structural solution strategies under element constraint for incremental damage. Computer Methods in Applied Mechanics and Engineering, 90(1), 869–884
Engineering Multi-Agent Systems and Generative AI: Report from the Agent Toolkits 2025 Community Session
No full textInternational audienceIntelligent agents have been a cornerstone of Artificial Intelligence (AI) since its early days-and received significant attention in the 1990s when the notion of autonomous agent was established, giving rise to research on Engineering Multi-Agent Systems (EMAS). Traditionally, this area has focused on theories, architectures, methodologies, paradigms, and languages for designing, implementing, and governing systems of autonomous agents. More recently, advances in Generative AI-and specifically large language models-have led to a new generation of agents and multi-agent systems, often referred to as Agentic AI.</div
Propriétés et comportements mécaniques de matériaux cimentaires : estimation par modélisations chimique et micromécanique
No full textInternational audiencePropriétés et comportements mécaniques de matériaux cimentaires : estimation par modélisations chimique et micromécaniqu
Agricultures machiniques: Représentation du travail agricole mécanisé sur YouTube
No full textNational audienceAgricultures machiniquesReprésentation du travail agricole mécanisé sur YouTube