Portail HAL edf
Not a member yet
11029 research outputs found
Sort by
Réponse dynamique d'une structure sur Inclusions Rigides : approche multiéchelle / Table ronde
No full textdisponible sur https://www.youtube.com/watch?v=3lUdMT9uQqg&list=PLwDLuT_SpnNTrmoV61iQajdWXFxnNfdFp&index=5&t=32sDeux temps forts composent cet enregistrement :1) La session "Réponse dynamique d'une structure sur Inclusions Rigides : approche multiéchelle", Présidée par Donatienne LEPAROUX (Université Gustave Eiffel) Introduction (Jesùs PEREZ, Terrasol) Enseignements sur la réponse cinématique et inertielle des fondations sur IR (Yuxiang SHEN, Terrasol) Développement d'une approche de type macroélément et validation à partir d'essais en centrifugeuse (Zheng LI, Université Gustave Eiffel) Simulation numérique des essais en centrifugeuse (Charbel NOHRA, Université Gustave Eiffel) Prise en compte des effets de site dans l'étude de la réponse des fondations sur IR (Fernando LOPEZ-CABALLERO, CentraleSupelec)2) Conclusion générale sous forme d'une table ronde (animée par Jesús PÉREZ-HERREROS, Terrasol et Luc THOREL, Université Gustave Eiffel), avec Fahd CUIRA (Terrasol), Cyril SIMON (EDF), Jérôme RACINAIS (Ménard), Hubert GIRAUD (SNCF Réseau)Cette vidéo a été captée le 17 décembre 2024 à Aix-en-Provence sur le Campus Méditerranée du Cerema, lors de la journée de clôture du projet ANR ASIRIplus-SDS : Amélioration des Sols par Inclusions Rigides, Sollicitations Dynamiques et Sismiques (https://anr.fr/Projet-ANR-19-CE22-0015)L'ensemble de la journée a été filmée. Les vidéos sont disponibles sur https://www.youtube.com/playlist?list=PLwDLuT_SpnNTrmoV61iQajdWXFxnNfdFp, mais aussi sur HAL : https://hal.science/hal-05426459v1, https://hal.science/hal-05520272v1, https://hal.science/hal-05520383v1, https://hal.science/hal-05520223, https://hal.science/hal-05426490v
Automated Deep Learning for Load Forecasting
No full textInternational audienceAccurate forecasting of electricity consumption is essential to ensure the performance and stability of the grid, especially as the use of renewable energy increases. Forecasting electricity is challenging because it depends on many external factors, such as weather and calendar variables. While regression-based models are currently effective, the emergence of new explanatory variables and the need to refine the temporality of the signals to be forecasted is encouraging the exploration of novel methodologies, in particular deep learning models. However, Deep Neural Networks (DNNs) struggle with this task due to the lack of data points and the different types of explanatory variables (e.g. integer, float, or categorical). In this paper, we explain why and how we used Automated Deep Learning (AutoDL) to find performing DNNs for load forecasting. We ended up creating an AutoDL framework called EnergyDragon by extending the DRAGON package and applying it to load forecasting. EnergyDragon automatically selects the features embedded in the DNN training in an innovative way and optimizes the architecture and the hyperparameters of the networks. We demonstrate on the French load signal that EnergyDragon can find original DNNs that outperform state-of-the-art load forecasting methods as well as other AutoDL approaches
Environmental cues and phenological variations in the spawning migration of the Allis shad (Alosa alosa) along the French Atlantic coast
No full textInternational audienc
Dynamique saisonnière des communautés de végétaux aquatiques dans différents contextes hydro-géomorphologiques du Rhône
No full textNational audienc
CO2 Capture in MDEA-PZ and AMP-PZ Solvents: ELECNRTL vs. ENRTL-RK Models and Their Efficiency in Natural Gas Combined Cycles
No full textInternational audienceThe necessity to mitigate global warming has intensified research into efficient CO2 capture processes, which are of crucialimportance to the carbon capture, utilization, and storage (CCUS) industry. Chemical absorption, particularly using aqueous amine blends, has emerged as a mature technology for post-combustion carbon capture due to its efficiency, energy requirements, and operational stability. This study aims to enhance the thermodynamic modelling of the CO2-PZ-MDEA-H2O system in Aspen Plus V12.1 software, employing ELECNRTL and ENRTL-RK models. The initial data collection and analysis involved 952 literature sources to optimize model’s thermodynamic parameters using the Maximum Likelihood method, resulting in significant accuracy improvements of 34% and 68% for ELECNRTL and ENRTL-RK, respectively. The process models were validated against data from the EDF pilot plant in Le Havre demonstrating their efficacy in simulating CO2 capture processes. Furthermore, simulations within Aspen Plus were conducted to evaluate the performance of various packing types, liquid holdup correlations, mass transfer models, and flow configurations. This was deemed crucial for process and column design. The study also investigated the use of AMP-PZ blends, which had been previously validated by NETL at another European pilot plant, to determine their potential for CO2 capture applications. While AMP-PZ solvent requires a lower liquid gas ratio than MDEA-PZ for CO₂ capture, it consumes more heat; increasing solvent concentration and using flue gas recirculation reduces heat demand, but capturing over 99% CO₂ significantly increases heat consumption due to the need for very low lean solvent loading
Développement d'une méthode DEM polyédrique pour la simulation de la relocalisation du combustible nucléaire lors d'un APRP
No full textNational audienceLors d’un Accident de Perte de Réfrigérant Primaire (APRP), les crayons combustibles sont soumis à des sollicitations thérmo-mécaniques intenses. Celles-ci peuvent provoquer un ballonnement ponctuel de la gaine dans lequel le combustible, présent dans un état fragmenté, peut se relocaliser radialement et axialement et induire un état de température élevé conduisant à la rupture de la gaine, donc de la première barrière de sûreté. Des recherches visent à étudier cet effet grâce à des modèles numériques avancés. Ici, on propose d’assimiler le combustible fragmenté à un milieu granulaire et de mettre en œuvre une modélisation par éléments discrets de la relocalisation des fragments. L’objectif final étant de prendre en compte l’interaction entre fragments et gaz de fission ainsi que la rupture de la gaine et l’expulsion des fragments, un premier objectif est de reprendre la méthode des éléments discrets (DEM) disponible dans le code Europlexus, et de l’étendre à des géométries polyédriques, en s’inspirant d’une méthode utilisée dans le code Rockable basée sur des sphéro-polyèdres
Direct parametrisation of invariant manifolds for non-autonomous forced systems including superharmonic resonances
No full textInternational audienceThe direct parametrisation method for invariant manifold is a model-order reduction technique that can be applied to nonlinear systems described by PDEs and discretised e.g. witha finite element procedure in order to derive efficient reduced-order models (ROMs). In non-linear vibrations, it has already been applied to autonomous and non-autonomous problemsto propose ROMs that can compute backbone and frequency-response curves of structures with geometric nonlinearity. While previous developments used a first-order expansion tocope with the non-autonomous term, this assumption is here relaxed by proposing a different treatment. The key idea is to enlarge the dimension of the parametrising coordinateswith additional entries related to the forcing. A new algorithm is derived with this starting assumption and, as a key consequence, the resonance relationships appearing throughthe homological equations involve multiple occurrences of the forcing frequency, showing that with this new development, ROMs for systems exhibiting a superharmonic resonance,can be derived. The method is implemented and validated on academic test cases involving beams and arches. It is numerically demonstrated that the method generates efficient ROMsfor problems involving 3:1 and 2:1 superharmonic resonances, as well as converged results for systems where the first-order truncation on the non-autonomous term showed a clear limitation
Light models intercomparison in agriVoltaics context
No full textSource Agritrop Cirad (https://agritrop.cirad.fr/612326/) * Autres projets (id;sigle;titre): ;ADELI;(FRA) ADELI//International audienceReduction of greenhouse gas emissions is one of the major issues facing our society when the population of the world is set to rise to 9.1 billion by 2050, 34 percent more than today. In this context, research is being conducted to develop new efficient, resilient and sustainable agricultural production systems based on agroecology but also integrating agrivoltaism concept. If these Agrivoltaics systems (agri-PV) can be efficient for some productions in favorable soil and climatic contexts, solutions still need to be developed to guarantee synergy between energy production and agronomic yield for many markets gardening, cereal or arboriculture crops as well as for livestock. Light is one of the major variables that needs to be characterised in order to analyse the conditions of synergy, as it is the main environmental factor involved in Photovoltaic (PV) system functioning and in the biophysical processes of crops, such as photosynthesis, evapotranspiration and photomorphogenesis, which are involved in plant growth and development. Because of climate variations from one year to the next [1], experiments should be conducted over several years in order to draw some conclusions. To address such problem, we propose to share the different datasets acquired during the various agri-PV experiments to allow robust statistical analysis. Moreover, analysis of light availability is not only based on measurements but also on the use of light models. These models are based on the same approach in both cases for the PV systems and the crops. Therefore, it exists different models (for instance based on raytracing or radiosity methods) that rely on similar assumptions regarding input data such as the characterisation of the concerned PV system and the physical modelling of incident irradiance (e.g., diffuse and direct radiation computation). As part of a collaborative approach of data sharing to analyse the conditions for synergy between the two productions, it becomes determinant do evaluate the precision of the different models. This kind of intercomparison was done in the research topics of remote sensing with the RAMI project [2] or in plant ecophysiology [3]. In these examples, no measurements were used because one model was considered as a reference model. In the present study, the objective is to quantify the precision of light models by simulating light below different configurations of agri-PV systems and compare them with actual measurements. The comparison will highlight the differences in physical modelling assumptions between the models and will include an estimate of biases and errors, as well as efficiency in terms of computational time
Functional diagnosis of industrial soils: from a cognitive model to in situ implementation
No full textInternational audienceIndustrial activities, such as thermal power plants, induce soil degradation on large areas (e.g. soil sealing, contamination related to fuel, coal and ash deposits, soil compaction). After the cessation of activities, landowners of such sites have a huge land heritage that could be considered to promote rehabilitation projects for new land-uses in the frame of the No Net Land Take by 2050. Therefore, there is a need to develop a robust and easy-to-use approach for landowners that could be implemented by soil techniciens/pratitioners to assess soil functions to measure their potential for future uses.First a cognitive model linking soil functions to a minimum dataset of indicators was established based on chemical, physical and biological properties of soil as well as vegetation cover. This cognitive model includes 6 soil functions (e.g. plant biomass production) and 17 sub-functions (e.g. phytoavailability of nutrients, nutrients storage) and a minimum set of indicators selected among a large list from research studies and attributed to each sub-function and function.Then two thermal power plants under closure were selected and a documentary survey was carried out for each site to identify contrasted zones in terms of soil cover, mostly based on the nature of the past activities (e.g. coal, slag or ash deposit, building foundations, fuel storage). Twelve zones considered as homogeneous in terms of vegetation and soil type and distinct from each other were selected on these two sites. In total, 12 soil profiles and 164 soil samples were analysed for various biological (plants, nematodes, microbial communities), chemical and physical parameters.Our results show contrasting situations. Despite the high vegetation cover of the three different ash deposit zones, their plant diversity indices ranged from very low to medium. The same goes for the area where building foundations were located, but they had very little vegetation cover. Also, the enrichment index (EI) and structure index (SI) of the nematode community showed that ash deposits are degraded, nutrient-poor soils and have a high C/N (>12) while the building foundation areas have a "mature and fertile" soil with optimal C/N. Whereas some soils could be considered as natural references as they were not affected by industrial activities, others were Technosols made of 100% artefacts. However, the gradient of anthropisation was surprisingly not correlated to the level of functions that were assessed. As an example, technogenic soils developed from fly ash exhibit high soil functions ratings (e.g. carbon storage).These initial results suggest that the functioning of these soils must be evaluated according to different scales (e.g. plot scale, surface soils), points of view (biological, chemical and physical) and soil functions (e.g. storage and sequestration of GHG, biodiversity reservoir), to establish their functional profiles and suggest possible future uses
Observer et explorer pour accompagner les grimpeurs vers leur sommet de performance
No full textInternational audiencePhysique, gestuel et mental sont, pour Cécile Avezou, les trois versants de la performance en escalade.Trois aspects qu’elle travaille en partant du grimpeur en tant qu’individu, tout en tenant compte de l’ensemble des paramètres de la discipline sportive, de la compétition et du profil du sportif. Une approche individualisée et coconstruite qui se fonde à la fois sur l’autonomie de l’athlète et la dynamique du collectif pour favoriser l’adaptation et réduire l’écart entre compétition et entraînement