Portail HAL edf
Not a member yet
11029 research outputs found
Sort by
Efficacité des restaurations du Vieux Rhin (France) : une analyse rétrospective d’une décennie de chroniques de suivi des communautés piscicoles
No full textInternational audienceRestoration of highly anthropized river environments represents a major challenge due to the profound alterations in these ecosystems and their associated processes. Restoration actions on the Old Rhine (Upper Rhine, France) have been implemented since 2013 to improve aquatic habitat quality and support associated biodiversity. This study evaluates the effectiveness of three types of intervention over a period of up to 12 years post-implementation, using a BACI protocol and monitoring more than 20,000 data points on fish communities and aquatic habitats. The interventions studied include gravel augmentation and controlled bank erosion, with or without the installation of artificial transverse groynes. Our results showed a diversification of aquatic habitats within the restored areas; however, this diversification remains localized to the intervention zones, with no significant downstream propagation. Certain types of intervention created unprecedented conditions in the river, fostering the emergence of hydraulic and thermal aquatic refuge zones that are favorable to juvenile fish. The impact of invasive alien species on the structure of fish communities was also discussed. Ultimately, the results provided insights into a crucial question regarding the effectiveness and durability of restoration actions in these heavily anthropized environments, with strong potential for application to other river systems.La restauration d'environnements fluviaux fortement anthropisés représente un défi majeur, en raison des altérations profondes de ces écosystèmes et des processus associés. Les actions de restauration du Vieux Rhin (Rhin Supérieur, France) ont été mises en œuvre à partir de 2013 pour améliorer la qualité des habitats aquatiques et soutenir la biodiversité associée. Cette étude évalue l’efficacité de 3 types d’intervention sur une période allant jusqu'à 12 ans post-travaux, au sein d’un protocole BACI avec une chronique de suivi de plus de 20 000 données concernant les communautés piscicoles et les habitats aquatiques. Les actions étudiées incluent des opérations par injections sédimentaires ou par érosion maîtrisée avec ou sans l’installation d’épis transversaux artificiels. Nos résultats montrent une diversification des habitats aquatiques dans les zones restaurées, mais cette diversification reste localisée à la zone de travaux sans diffusion significative vers l’aval. Certains types d’action ont favorisé des conditions inédites dans le fleuve, contribuant à l’émergence de zones refuges aquatiques hydrauliques et thermiques favorables à la présence de juvéniles. L'impact des espèces exotiques envahissantes sur la structure des communautés piscicoles est également discutée. In fine, les résultats apportent des éléments de réponse à une question cruciale sur l’efficacité et la durabilité des actions de restauration dans ces environnements fortement anthropisés, avec un fort potentiel d’application à d’autres systèmes fluviaux
Vous prendrez bien une tasse de thé? Routinisation et temporalité de la recharge du véhicule électrique : une sociologie des pratiques au prisme de trois terrains d’enquête
No full textInternational audienc
Condition monitoring and diagnostic of hydropower units based on machine learning techniques
No full textInternational audienceAs the energy sector accelerates its transition toward incorporating a greater proportion of renewable and low-carbon sources in the electricity mix, ensuring the availability of hydroelectric power plants is more crucial than ever to maintain the equilibrium between electricity supply and demand. This is particularly significant in Europe, where the ageing of power plant infrastructures raises concerns about the energy grid's ability to accommodate greater capacity volatility in the context of unstable and limited remaining hydro potential. For this reason, this paper introduces a diagnostic methodology based on machine learning techniques for detecting and characterising anomalous behaviours of hydro generators related to long-term degradation in the context of seasonal periodicity.The methodology leverages time-series of monitoring data acquired in an industrial context from hydro generators commercially operated by France's state-owned energy company "Électricité de France Hydro Sud Ouest". It employs dimensionality reduction techniques and clustering algorithms to semi-automatically identify anomalous changes in the machine production cycle, given its dataset specifications. The objective is to create an ensemble classifier composed of autoencoder-type neural networks trained to approximate the distribution of classes validated by human experts. The soft voting module based on the approximation errors of each autoencoder models is employed to create a proximity index to each class identified.The methodology aims at tackling the challenges of estimating the hydropower plant health state in the context of the sparsity of failure data and difficulty quantified effects of both seasonality and long-term degradations. Indeed, those long-lived multivariate systems feature unique designs that are adapted to their operational environments. Thus, the methodology proposes a unified framework which employ machine learning techniques to better adapt monitoring practice to the variety in system architecture while accommodate future integration of components and sensors. Therefore, it enhances operators understanding of monitoring data and generated alarms through there contextualisation while promoting a proactive strategy to discover anomalies in the system's behaviour. Consequently, it aims to support timely scheduled productive maintenance interventions
Dam-break flow over various obstacles configurations
No full textInternational audienceFast floods resulting from the failure of hydraulic structures can be characterized by ‘dam-break’ type waves. They pose catastrophic risks to downstream populations and result in severe structural damage, especially in urban areas. To assess and mitigate these risks, it is essential to forecast the influence of urban forms on flooding severity at a global scale. This paper provides datasets from reduced-scale physical experiments of transient flow through various obstacles configurations. The experiments are conducted in a rectangular horizontal open channel, where flow conditions are achieved by rapidly opening a gate holding a volume of water. To assess the impact of obstacle configurations on flow behaviour, two obstacle sizes are investigated, along with one idealized city layout. The experiments provide complete water hydrographs upstream and downstream of the gate. Additionally, the good performance of the code_saturne computational fluid dynamics (CFD) solver and the volume-of-fluid (VOF) method in numerically simulating the experiments is demonstrated
airGRteaching: Teaching Hydrological Modelling with the GR Rainfall-Runoff Models ('Shiny' Interface Included). R package version 0.3.5
No full textAdd-on package to the 'airGR' package that simplifies its use and is aimed at being used for teaching hydrology. The package provides 1) three functions that allow to complete very simply a hydrological modelling exercise 2) plotting functions to help students to explore observed data and to interpret the results of calibration and simulation of the GR ('Génie rural') models 3) a 'Shiny' graphical interface that allows for displaying the impact of model parameters on hydrographs and models internal variables
De l'explicabilité à l'explication des IA : perspective située incarnée par et pour les métiers
No full textInternational audienceArtificial intelligence (AI) has made great strides, thanks in particular to machine learning algorithms, but the increasing complexity of models poses a problem of transparency. Explainable Artificial Intelligence (XAI) has been proposed to make AI more transparent and foster its appropriation. Although numerous studies have identified challenges and avenues for research, these remain highly technocentric. This article is part of a thesis on the explicability and appropriation of artificial intelligence (AI) in professional and socio-domestic environments, focusing first on the professional environment. It aims to clarify the notion of explicability and provide a complementary, anthropocentric perspective to existing ones, based on an exploratory study in a professional context.L'intelligence artificielle (IA) a réalisé des progrès importants, notamment grâce aux algorithmes d'apprentissage automatique, mais l'augmentation de la complexité des modèles pose un problème de transparence. L'intelligence artificielle explicable (XAI) a été proposée pour rendre l'IA plus transparente et favoriser son appropriation. Bien que de nombreuses études aient identifié des défis et des pistes de recherche, celles-ci restent très technocentrées. Cet article s'inscrit dans le cadre d'une thèse sur l'explicabilité et l'appropriation de l'intelligence artificielle (IA) dans les environnements professionnels et socio-domestiques, en se concentrant d'abord sur le milieu professionnel. Il vise à clarifier la notion d'explicabilité et à apporter une perspective complémentaire et anthropocentrée aux perspectives existantes, en s'appuyant sur une étude exploratoire en contexte professionnel
Heat Stress assessment using an urban microclimate zonal model at the block scale coupled with building models
No full textInternational audienc
SpinDoctor-IVIM: A virtual imaging framework for intravoxel incoherent motion MRI
No full textInternational audienceIntravoxel incoherent motion (IVIM) imaging is increasingly recognised as an important tool in clinical MRI, where tissue perfusion and diffusion information can aid disease diagnosis, monitoring of patient recovery, and treatment outcome assessment. Currently, the discovery of biomarkers based on IVIM imaging, similar to other medical imaging modalities, is dependent on long preclinical and clinical validation pathways to link observable markers derived from images with the underlying pathophysiological mechanisms. To speed up this process, virtual IVIM imaging is proposed. This approach provides an efficient virtual imaging tool to design, evaluate, and optimise novel approaches for IVIM imaging. In this work, virtual IVIM imaging is developed through a new finite element solver, SpinDoctor-IVIM, which extends SpinDoctor, a diffusion MRI simulation toolbox. SpinDoctor-IVIM simulates IVIM imaging signals by solving the generalised Bloch-Torrey partial differential equation. The input velocity to SpinDoctor-IVIM is computed using HemeLB, an established Lattice Boltzmann blood flow simulator. Contrary to previous approaches, SpinDoctor-IVIM accounts for volumetric microvasculature during blood flow simulations, incorporates diffusion phenomena in the intravascular space, and accounts for the permeability between the intravascular and extravascular spaces. The above-mentioned features of the proposed framework are illustrated with simulations on a realistic microvasculature model.</div
Phenotypic diversity of brown trout (Salmo trutta L.) during the late Upper Pleistocene and Early Holocene in the glacial refugium of Iberia
No full textInternational audienceHuman populations have relied on fish for food for thousands of years, as evidenced in numerous Palaeolithic archaeological sites across Europe. In this paper, we investigate past populations of brown trout (Salmo trutta L.) along the Bay of Biscay coast before, during, and after the Last Glacial Maximum and compare their phenotypic diversity with modern populations. A total of 270 salmonid vertebrae from 11 archaeological sites in Spain and France were analysed, which span c. 30 to 9 kyr cal B.P. Species identification, combining ZooMS analyses with vertebral body size, confirmed 219 specimens as brown trout, and we present a model for salmonid species identification based on vertebral size. Stable isotope analyses of carbon and nitrogen reveal a continuum of habitat use patterns. After testing multiple published regression models for body length estimation in brown trout, we develop a novel linear regression model alongside a protocol for age estimation. Overall, the analysis of body size, age and migratory ecotype demonstrates that the phenotypic diversity of brown trout populations during the Pleistocene-Holocene transition is comparable to that of modern populations, reinforcing the role of the Iberian glacial refugium in maintaining diversity during the Last Glacial Maximum. The finding of large freshwater trout alongside sea trout in human-occupied archaeological sites further suggests the exploitation of both riverine and marine fish resources by human populations.</div
Corrosion behaviour of Alloy 600 containing oxide inclusions exposed to primary water of pressurised water reactors
No full textInternational audienceIntergranular stress corrosion cracking (IGSCC) was observed in an Alloy 600 bottom-mounted instrumentation (BMI) nozzle of a French pressurised water reactor (PWR). Destructive examinations of the component revealed the presence of oxide inclusions extending along the crack, which were identified as likely to promote IGSCC initiation. In this context, the corrosion behaviour of Alloy 600 laboratory heats containing aluminium and/or magnesium oxide inclusions was investigated in primary water at 325°C and 360°C for oxidation times ranging from 10 h to 7000 h. Complete dissolution and/or detachment of the oxide inclusions was observed after a few hundred hours of exposure to primary water, and the presence of magnesium borate was locally detected. Furthermore, blister-like features consisting of Ni2FeBO5 filaments embedded in a double nickel chromite spinel layer were characterised on the surface of Alloy 600 in the vicinity of some alumina inclusions while, a micron-thick oxide layer was formed at the former interface with the pre-existing inclusions