HAL-INSA Toulouse
Not a member yet
34325 research outputs found
Sort by
Reconstruction of flow fields from data using physics-informed gaussian process regression
No full textInternational audienc
Certification Gap Analysis for Normal-Category and Large Hydrogen-Powered Airplanes
No full textInternational audienceThe transition to hydrogen as an aviation fuel, as outlined in current decarbonization roadmaps, is expected to result in the entry into service of hydrogen-powered aircraft in 2035. To achieve this evolution, certification regulations are key enablers. Due to the disruptive nature of hydrogen aircraft technologies and their associated hazards, it is essential to assess the maturity of the existing regulatory framework for certification to ensure its availability when manufacturers apply for aircraft certification. This paper presents the work conducted under the Clean Aviation CONCERTO project to advance certification readiness by comprehensively identifying gaps in the current European regulations. Generic methodologies were developed for regulatory gap and risk analyses and applied to a hydrogen turbine aircraft with non-propulsive fuel cells as the APU. The gap analysis, conducted on certification specifications for large and normal-category airplanes as well as engines, confirmed the overall adequacy of many existing requirements. However, important gaps exist to appropriately address hydrogen hazards particularly concerning fire and explosion, hydrogen storage and fuel systems, crashworthiness, and occupant survivability. The paper concludes by identifying critical areas for certification and highlighting the need for complementary hydrogen phenomenology data, which are key to guiding future research and regulatory efforts for certification readiness maturation.<br /
Evolution of Measures in Nonsmooth Dynamical Systems: Formalisms and Computation
No full textInternational audienceThis article develops mathematical formalisms and provides numerical methods for studying the evolution of measures in nonsmooth dynamical systems using the continuity equation. The nonsmooth dynamical system is described by an evolution variational inequality and we derive the continuity equation associated with this system class using three different formalisms. The first formalism consists of using the {superposition principle} to describe the continuity equation for a measure that disintegrates into a probability measure supported on the set of vector fields and another measure representing the distribution of system trajectories at each time instant. The second formalism is based on the regularization of the nonsmooth vector field and describing the measure as the limit of a sequence of measures associated with the regularization parameter. In doing so, we obtain quantitative bounds on the Wasserstein metric between measure solutions of the regularized vector field and the limiting measure associated with the nonsmooth vector field. The third formalism uses a time-stepping algorithm to model a time-discretized evolution of the measures and show that the absolutely continuous trajectories associated with the continuity equation are recovered in the limit as the sampling time goes to zero. We also validate each formalism with numerical examples. For the first formalism, we use polynomial optimization techniques and the moment-SOS hierarchy to obtain approximate moments of the measures. For the second formalism, we illustrate the bounds on the Wasserstein metric for an academic example for which the closed-form expression of the Wasserstein metric can be calculated. For the third formalism, we illustrate the time-stepping based algorithm for measure evolution on an example that shows the effect of the concentration of measures
High-dimensional Parameter Identification with Physics-Informed Machine Learning for Flood Prediction
No full textInternational audienceVariational Data Assimilation (VDA) frameworks, particularly those using the adjoint technique, have long been the preferred approach for calibrating physical model parameters to align with observed data [1]. However, these methods may appear complex and still computationally expensive for high-dimensional parameter identification. Recent advances in Machine Learning, especially Physics-Informed Neural Networks (PINNs), present exciting opportunities to address these challenges [2].This work revisits two PINN-based approaches for inverse problems, emphasizing their ability to infer high-dimensional physical parameters. The first, termed here Fully-Parameterized PINN, constructs a parameter-differentiable surrogate model through initial offline training, followed by rapid online parameter identification. This method treats physical parameters as NN inputs, making it prone to the curse of dimensionality. The second variant, called here Semi-Parameterized PINN (SP-PINN), integrates physical parameters as NN parameters, enabling efficient inference regardless of dimensionality via automatic differentiation.In this work, these methods are applied to the inference of spatially-distributed physical parameters (e.g., friction or infiltration coefficients) in flood models, a critical task for improving forecast accuracy. To evaluate their performance, several numerical experiments will be presented, including cases based on real-world data. In particular, SP-PINN is tested on a representative scenario for identifying a 1000-dimensional spatial friction parameter in a Shallow-Water model [3]. Comparisons with more traditional VDA methods will also be shown, demonstrating the simplicity and efficiency of SP-PINN, and thus establishing it as a viable alternative in real-world parameter identification tasks.REFERENCES[1]Monnier, J. (2021). Data Assimilation. Inverse Problems, Assimilation, Control, Learning. Lectures notes. URL: https://hal.science/hal-03040047.[2]Tanyu, D. N., Ning, J., Freudenberg, T., Heilenkötter, N., Rademacher, A., Iben, U., & Maass, P. (2023). Deep learning methods for partial differential equations and related parameter identification problems. Inverse Problems, 39(10), 103001.[3]Boulenc, H., Bouclier, R., Garambois, P. A., & Monnier, J. (2025). Spatially-Distributed Parameter Identification by Physics-Informed Neural Networks illustrated on the Shallow-Water Equations. Preprint (to appear in Inverse Problems)
Opportunities and Challenges towards Portable Nanotechnology Processes over Open Facilities
No full textInternational audienc
Obtuse almost-equiangular sets
No full text29 pages; fixed problem with references from previous versionFor , a set of points on the -dimensional unit sphere is called -almost equiangular if among any three distinct points there is a pair with inner product . We propose a semidefinite programming upper bound for the maximum cardinality of such a set based on an extension of the Lov\'asz theta number to hypergraphs. This bound is at least as good as previously known bounds and for many values of and it is better. We also refine existing spectral methods to show that for all and , with equality only at . This allows us to show the uniqueness of the optimal construction at for and to enumerate all possible constructions for and
Luigi Broglio (biografia breve)
No full texthttps://disf.org/autori/luigi-broglioLuigi Broglio was an Italian aerospace engineer, university professor, and visionary behind Italy’s entry into the space age. Often called the "father of Italian space exploration," he played a leading role in launching San Marco 1 in 1964—Italy’s first satellite—making it the third country to launch a satellite using its own facilities.A former professor at the University of Rome and an expert in aerodynamics, Broglio collaborated with NASA to develop the San Marco Project, which included a mobile sea-based launch platform off the coast of Kenya. Under his leadership, Italy developed several scientific satellites and gained early access to space-based research.Broglio's legacy includes the Luigi Broglio Space Center in Kenya, which has had a lasting impact on international space cooperation.Luigi Broglio était un ingénieur aérospatial italien, professeur d'université et visionnaire à l'origine de l'entrée de l'Italie dans l'ère spatiale. Souvent surnommé le « père de l'exploration spatiale italienne », il a joué un rôle majeur dans le lancement de San Marco 1 en 1964, le premier satellite italien, devenant ainsi le troisième pays à lancer un satellite avec ses propres installations.Ancien professeur à l'Université de Rome et expert en aérodynamique, Broglio a collaboré avec la NASA pour développer le projet San Marco, qui comprenait une plateforme de lancement mobile en mer au large des côtes kenyanes. Sous sa direction, l'Italie a développé plusieurs satellites scientifiques et a bénéficié d'un accès précoce à la recherche spatiale.L'héritage de Broglio comprend le Centre spatial Luigi Broglio au Kenya, qui a eu un impact durable sur la coopération spatiale internationale.Luigi Broglio è stato un ingegnere aerospaziale italiano, professore universitario e visionario dietro l'ingresso dell'Italia nell'era spaziale. Spesso definito il "padre dell'esplorazione spaziale italiana", ha avuto un ruolo di primo piano nel lancio del San Marco 1 nel 1964, il primo satellite italiano, rendendo l'Italia il terzo Paese a lanciare un satellite utilizzando le proprie strutture.Ex professore all'Università di Roma ed esperto di aerodinamica, Broglio ha collaborato con la NASA allo sviluppo del Progetto San Marco, che includeva una piattaforma di lancio mobile marittima al largo delle coste del Kenya. Sotto la sua guida, l'Italia ha sviluppato diversi satelliti scientifici e ha ottenuto un accesso precoce alla ricerca spaziale.L'eredità di Broglio include il Centro Spaziale Luigi Broglio in Kenya, che ha avuto un impatto duraturo sulla cooperazione spaziale internazionale
Conception et réalisation d'un système de détection de chute avant impact au sol : IA embarquée et micro-actionneur pyrotechnique
No full textNational audienceThis multi-disciplinary research project presents a low-power, portable, on-board system for detecting pre-falls in the elderly, in order to protect them by deploying an airbag. The original detection system requires only a single triaxial accelerometer and uses gas-generating energetic composites. In operation, it is an on-board AI (CNN or SNN) that decides whether or not to initiate combustion of the energetic material to generate the gas. This work has demonstrated the feasibility of a lightweight, miniature on-board system (in place of gas cylinders) to protect elderly people from falls, while increasing comfort.Ce travail de recherche multidisciplinaire présente un système embarqué portable de faible puissance dédié à la détection de pré chute des personnes âgées afin de les protéger grâce au déploiement d'un airbag. L'originalité du système de détection ne requiert qu'un seul accéléromètre triaxial et utilise des composites énergétiques générateurs de gaz. Dans fonctionnement, c'est une IA embarquée (CNN ou SNN) qui décide d'initier ou pas la combustion du matériau énergétique pour générer le gaz. Ces travaux ont démontré la faisabilité d'un système embarqué léger et miniature (à la place des bouteilles de gaz) pour protéger les personnes âgées des chutes tout en augmentant ainsi le confort
Translaminar fracture behavior of hybrid woven-ply peek thermoplastic laminates under isothermal and kerosene flame exposure
No full textInternational audienceThe growing demand to use thermoplastic matrix composites in aeronautic has been confronted with the need to understand the fracture mechanisms under different service conditions. Thus, the present work aims at studying the influence of different thermal stress conditions on the fracture behavior of PolyEther Ether Ketone-reinforced carbon/glass fiber hybrid laminates (CG/PEEK), as a function of heating temperature and fire exposure time. Compact Tensile (CT) specimens are exposed to isothermal (from 350℃ to 550℃ in a high temperature furnace) and critical service conditions (kerosene flame exposure characterized by a heat flux of 116 kW/m2 and a temperature of 1150℃) exposure conditions. Monotonic tensile tests are then conducted to assess the mode I translaminar fracture toughness (FT) at room temperature. Crack propagation is monitored during mechanical loading using a Digital Image Correlation (DIC) device combined with a binarization algorithm. Then, the G-R curves have been obtained from the compliance method. Under isothermal conditions, the residual mechanical properties degrade as temperature increases, particularly once thermal decomposition is about to start, due to the formation of porosities and extensive delamination. Under flame exposure, the microscopic and tomographic observations reveal thermally- and mechanically-induced damages with a heterogeneous distribution due to temperature gradients within the plies of laminates. Depending on the pyrolysis degree of each ply, the load bearing capabilities of the plies gradually deteriorate from the exposed to the opposed side. The critical FT values shows 32 a decreasing trend (as a function of fire exposure time) with increasing ply number of charred regions through the thickness
Exciton Self-Trapping in Twisted Hexagonal Boron Nitride homostructures
No full textInternational audienc