HAL Université de Toulouse, et Toulouse INP
Not a member yet
159308 research outputs found
Sort by
Optimisation bayésienne d'un réseau neuronal léger et précis pour la prédiction des performances aérodynamiques
No full textInternational audienceEnsuring high accuracy and efficiency of predictive models is paramount in the aerospace industry, particularly in the context of multidisciplinary design and optimization processes. These processes often require numerous evaluations of complex objective functions, which can be computationally expensive and time-consuming. To build efficient and accurate predictive models, we propose a new approach that leverages Bayesian Optimization (BO) to optimize the hyper-parameters of a lightweight and accurate Neural Network (NN) for aerodynamic performance prediction. To clearly describe the interplay between design variables, hierarchical and categorical kernels are used in the BO formulation. We demonstrate the efficiency of our approach through two comprehensive case studies, where the optimized NN significantly outperforms baseline models and other publicly available NNs in terms of accuracy and parameter efficiency. For the drag coefficient prediction task, the Mean Absolute Percentage Error (MAPE) of our optimized model drops from 0.1433% to 0.0163%, which is nearly an order of magnitude improvement over the baseline model. Additionally, our model achieves a MAPE of 0.82% on a benchmark aircraft self-noise prediction problem, significantly outperforming existing models (where their MAPE values are around 2 to 3%) while requiring less computational resources. The results highlight the potential of our framework to enhance the scalability and performance of NNs in large-scale MDO problems, offering a promising solution for the aerospace industry.Garantir la précision et l'efficacité des modèles prédictifs est primordial dans l'industrie aérospatiale, en particulier dans le contexte des processus de conception et d'optimisation multidisciplinaires. Ces processus nécessitent souvent de nombreuses évaluations de fonctions objectives complexes, ce qui peut être coûteux en temps et en argent. Pour construire des modèles prédictifs efficaces et précis, nous proposons une nouvelle approche qui s'appuie sur l'optimisation bayésienne (BO) pour optimiser les hyperparamètres d'un réseau neuronal (NN) léger et précis pour la prédiction des performances aérodynamiques. Pour décrire clairement l'interaction entre les variables de conception, des noyaux hiérarchiques et catégoriels sont utilisés dans la formulation BO. Nous démontrons l'efficacité de notre approche à travers deux études de cas complètes, où le réseau neuronal optimisé surpasse de manière significative les modèles de base et d'autres réseaux neuronaux publiquement disponibles en termes de précision et d'efficacité des paramètres. Pour la tâche de prédiction du coefficient de traînée, l'erreur absolue moyenne en pourcentage (MAPE) de notre modèle optimisé passe de 0,1433 % à 0,0163 %, ce qui représente une amélioration de près d'un ordre de grandeur par rapport au modèle de base. En outre, notre modèle atteint un MAPE de 0,82 % sur un problème de référence de prédiction du bruit propre d'un avion, ce qui est nettement supérieur aux modèles existants (dont les valeurs MAPE sont de l'ordre de 2 à 3 %) tout en nécessitant moins de ressources informatiques. Les résultats soulignent le potentiel de notre cadre pour améliorer l'évolutivité et la performance des NN dans les MDO à grande échelle
Monopole excitations in the U(1) Dirac spin liquid on the triangular lattice
No full textInternational audienceThe U(1) Dirac spin liquid might realize an exotic phase of matter whose low-energy properties are described by quantum electrodynamics in 2+1 dimensions, where gapless modes exist but spinons and gauge fields are strongly coupled. Its existence was proposed in frustrated Heisenberg models in the presence of frustrating superexchange interactions by the (Abrikosov) fermionic representation of the spin operators [Wen, Phys. Rev. B 65, 165113 (2002)], supplemented by the Gutzwiller projection. Here, we construct charge-Q monopole excitations in the Heisenberg model on the triangular lattice with nearest-neighbor (J1) and next-nearest-neighbor (J2) couplings. In the highly frustrated regime, singlet and triplet monopoles with Q=1 become gapless in the thermodynamic limit; in addition, the energies for generic Q agree with field-theoretical predictions, obtained for a large number of gapless fermion modes. Finally, we consider localized gauge excitations, in which magnetic π fluxes are concentrated in the triangular plaquettes (in analogy with Z2 visons), showing that these kinds of states do not play a relevant role at low energies. All our findings lend support to a stable U(1) Dirac spin liquid in the J1-J2 Heisenberg model on the triangular lattice
Influence of stress variation on radar wave propagation in concrete: Application to the monitoring of nuclear containment building
No full textInternational audienceContainment buildings of nuclear power plants are made of prestressed concrete. Maintaining the tension in prestressing tendons is essential for the safety of these structures. However, steel and concrete can be subject to ageing and pathology (creep, shrinkage, corrosion…), this can lead to prestressing losses. It is then crucial to be able to know the tension state of tendons. Current techniques used to evaluate tension in cables are semi-destructive and are then unusable on containment buildings. This is why it is relevant to develop non-destructive techniques to evaluate potential prestressing losses. Based on the inaccessibility of the cables, this study suggests evaluating concrete stress rather than cable tension. This work particularly focuses on the influence of stress on radar wave propagation in concrete. Tests show that a compressive stress increase on concrete slabs leads to an isotropic delay in the electromagnetic signal and an anisotropic decrease in amplitude, more pronounced in the direction normal to the stress direction. These variations highly depend on concrete hydric state, as it is observed that an increase in stress on a near-saturation or an oven-dried concrete does not induce a signal variation. Finally, in-situ tests performed onto a mock-up of a nuclear containment building built by EDF confirm the sensitivity of radar waves to a variation of concrete stress
Self-healing performance of thermally damaged ultra-high performance concrete: Rehydration and recovery mechanism
No full textInternational audienceConcrete suffers significant performance degradation when exposed to high temperatures. This study explored the beneficial role of waste glass powder (WGP) in mitigating thermal damage and ultra-high performance concrete (UHPC) after elevated temperature exposure. The mechanism was elucidated through the chemical and microstructure changes, the composition of hydrates after exposure to elevated temperatures, and the subsequent re-curing. The presence of WGP significantly enhanced the residual mechanical properties of UHPC due to more wollastonite generation. The WGP also facilitated the recovery of mechanical properties and surface morphology during the post-fire self-healing process. The microstructural results confirmed that the WGP promoted the formation of the wollastonite phase in the thermal-damaged UHPC by reacting with the dehydrated products. Thermodynamic simulations indicated that the incorporation of WGP in UHPC resulted in an increase of liquid phase and its early appearance at high temperatures led to the transformation of γ-C2S into more stable wollastonite phases. Meanwhile, the activation of unreacted WGP by limewater further generated secondary hydration products to reduce matrix porosity. These hydrates mainly consisted of C-(N)-S-H gels with a low calcium-to-silicon ratio (Ca/Si) and high sodium-to-silicon ratio (Na/Si) ratio, which could effectively fill the micropores and microcracks in UHPC. As a result, the densified microstructure induced by these regenerated C-(N)-S-H gels largely contributed to the recovery of the thermally damaged UHPC. The outcome of this study provides a decarbonization solution to address damages of UHPC exposed to fire conditions
Une proposition de loi pour le foot ou une proposition de loi pour le sport professionnel ? Patrick Bayeux
No full texthttps://patrickbayeux.com/actualites/une-proposition-de-loi-pour-le-foot-ou-une-proposition-de-loi-pour-le-sport-professionnel-patrick-bayeux/Si le football connait une crise profonde, peut on se satisfaire d'une loi spécifique qui vise à réparer un modèle dépassé ? Entre l’incertitude des droits TV dans un contexte de digitalisation croissance des services, l’illégalité du soutien financier accordé par les collectivités territoriales aux sports de salle, la difficulté de mettre à disposition une enceinte sportive sans mettre en concurrence les clubs, la difficulté de passer d’un modèle public privé à un modèle privé public, une réforme globale est impérative. Des états généraux du sport s'imposent #EGS202
L'expertise citoyenne dans la gouvernance de la catastrophe de Fukushima Daiichi : une résistance par la science face à la politique de résilience
No full textInternational audienc
An updated non-intrusive, multi-scale, and flexible coupling interface in WRF 4.6.0
No full textInternational audienceAbstract. The Weather Research and Forecasting (WRF) model has been widely used for various applications, especially for solving mesoscale atmospheric dynamics. Its high-order numerical schemes and nesting capability enable high spatial resolution. However, a growing number of applications are demanding more realistic simulations through the incorporation of coupling with new model compartments and an increase in the complexity of the processes considered in the model. (e.g., ocean, surface gravity wave, land-surface, chemistry...). The present paper details the development and the functionalities of the coupling interface we implemented in WRF. It uses the Ocean-Atmosphere-Sea-Ice-Soil – Model Coupling Toolkit (OASIS3-MCT) coupler, which has the advantage of being non-intrusive, efficient, and very flexible to use. OASIS3-MCT has already been implemented in many climate and regional models. This coupling interface is designed with the following baselines: (1) it is structured with a 2-level design through 2 modules: a general coupling module, and a coupler-specific module, allowing to easily add other couplers if required, (2) variables exchange, coupling frequency, and any potential time and grid transformations are controlled through an external text file, offering great flexibility, (3) the concepts of “external domains” and “coupling mask” are introduced to facilitate the exchange of fields to/from multiple sources (different models, fields from different models/grids/zooms...). Finally, two examples of applications of ocean-atmosphere coupling are proposed. The first is related to the impact of ocean surface current feedback to the atmospheric boundary layer, and the second concerns the coupling of surface gravity waves with the atmospheric surface layer
Measurement of Electrolyte Self-Diffusion in Laser Structured Electrodes
No full textInternational audienc
Depositional Environment of the Amapari Marker Band: Rising Water Levels Formed Kilometer‐Scale Lake in Gale Crater, Mars
No full textInternational audienceThe Amapari Marker Band (AMB) is a layer within the Mount Sharp stratigraphy that has been mapped around the Gale crater in orbital images and was recently investigated up close by the Curiosity rover. Symmetric wave ripple marks within the AMB indicate a lacustrine depositional environment in the area investigated along the Curiosity traverse. The wavelength and morphology of the ripples constrain the water depth to a few meters or less. The lateral continuity of the ripple unit defines a minimum extent of the lake during ripple formation. The stratigraphy of the AMB is consistent with an environment of increasing water depth during sedimentation and the lateral correlation of the AMB stratigraphy suggests a transgressive depositional system building upon an eroded surface. The location of the AMB within the surrounding aeolian stratigraphy, coupled with the progression of depositional environments through the Mirador formation, records a pattern of a rising water table relative to sedimentation rates. The potential regional extent of the lacustrine environment, based on orbital mapping of the AMB's variable elevation, spans at minimum 2.0 km of the lateral AMB deposit in the area around Marker Band Valley and may have extended up to 14 km to the west across the northern Gale crater
G+E copula model to improve the estimation of the genetic parameters in bivariate mixed model
No full textIn environmental sciences, especially in breeding, phenotypes are measured to improve traits of interest. These phenotypes can be decomposed into genetic and environmental components (G+E) and are therefore modelled using a mixed model. As the genetic part is unobservable, the breeding values are latent variables in the model, characterized by a covariance matrix associated with the system's pedigree. In most studies, multiple phenotypes are observed simultaneously, and their joint distribution is generally assumed to be Gaussian. Then, the estimates are obtained using a restricted maximum likelihood (REML) approach or a Bayesian inference, under Gaussian assumptions. However, even if each of the phenotypes is Gaussian, their joint distribution may not be due to a non-Normal dependence structure, which can be characterized using copula functions. When the joint distribution is atypical, such as in the case of heavy-tailed distributions, and the offspring arise from a selection process of the reproducers, the estimates of the variance components can be strongly biased. In this paper, we introduce a G+E mixed inference model, which generalizes the standard model used in genetics by incorporating copulas to account for various joint distributions of the phenotypes. We propose a stochastic gradient descent approach coupled with a Monte Carlo Markov Chain step to estimate the variance components and predict the genetic values. The performance of the algorithm is tested through simulations. Finally, the algorithm is applied on a true dataset of pig breeding, where the assumption of normality for the joint phenotype seems unrealistic. The estimates are compared with those obtained by REML under Gaussian assumptions.</div