Portail HAL ONERA
Not a member yet
12842 research outputs found
Sort by
Circuit-level simulation of TID and thermal regeneration effects on 28nm FDSOI technology
No full textInternational audienceIn this work, we implement physical models that describe the effects of radiation and thermal regeneration on 28nm FDSOI transistors. These models can be used with circuit simulation software to evaluate the radiation response in this technology. We perform SPICE simulations using UTSOI 2.0 from the transistor to the circuit level to determine the electrical degradation and the recovery by applying annealing cycles. In addition, we test a thermal-electrical mitigation strategy to partially or fully recover the TID-induced electrical degradation of the transistor
Out-of-Distribution Radar Detection in Compound Clutter and Thermal Noise through Variational Autoencoders
No full textInternational audienceThis paper presents a novel approach to radar target detection using Variational AutoEncoders (VAEs). Known for their ability to learn complex distributions and identify out-ofdistribution samples, the proposed VAE architecture effectively distinguishes radar targets from various noise types, including correlated Gaussian and compound Gaussian clutter, often combined with additive white Gaussian thermal noise. Simulation results demonstrate that the proposed VAE outperforms classical adaptive detectors such as the Matched Filter and the Normalized Matched Filter, especially in challenging noise conditions, highlighting its robustness and adaptability in radar applications.</div
Satellite Communication Resources Management in a Earth Observation Federation of Constellations
No full textInternational audienceThis paper addresses a new problem arising from the novel concept of multi-mission federation in Earth observation. This problem arises because of the development of several competing ground station networks that could be used by a component of federated missions called Satellite Communication and Resource Management System (SCRMS). Among the functions of SCRMS selection of contacts is a challenging decision problem. It is an optimization problem with Boolean variables corresponding to the selection of some contacts among potential contacts. Constraints enforce the fulfillment of the communication needs of the satellites. The considered criteria are the communication cost and the conflicts and jamming between satellites whencommunicating with sites, optimized lexicographically. We propose a linear program, several incomplete search schemes and greedy allocation schemes to address this problem, and evaluate them on realistic systems. Results indicate that the linear program complies with computation time requirements and produces better contact selection plans than other optimization schemes
Interpolation non linéaire de données CFD paramétrique d'écoulement compressible dominé par convection
No full textInternational audienceIn this study, we present a method to interpolate two parametric CFD fields computed from a high fidelity numerical scheme. This interpolation method can later be used to make cost-effective predictions for new parameters in CFD problems. The method relies on the computation of mappings during the offline phase to align the coherent structures of the flow. Once aligned, a convex interpolation is performed on the mapped data to produce a prediction. This method is computationally efficient in the offline phase as it only requires two evaluations of the full order model (FOM). Furthermore, the interpolation method is non-intrusive and can be applied to all type of CFD data fields. The method has been applied on a parametric problem NACA0012 airfoil. The prediction at a new parameter out perform the traditional linear models when compared to full order model simulations.Dans cette étude, nous présentons une méthode d’interpolation entre deux champs paramétriques CFD, calculés à l’aide d’un schéma numérique haute fidélité. Cette méthode d’interpolation peut ensuite être utilisée pour effectuer des prédictions à moindre coût pour de nouveaux paramètres. La méthode repose sur le calcul de champs de deformation en phase hors ligne afin d’aligner les structures cohérentes de l’écoulement. Une fois les champs alignés, une interpolation convexe est réalisée sur les données transformées pour produire une prédiction. Cette approche est efficace sur le plan computationnel lors de la phase hors ligne, car elle ne nécessite que deux évaluations du modèle haute fidélité (FOM). De plus, la méthode d’interpolation est non intrusive et peut être appliquée à tout type de champ de données CFD. La méthode a été testée sur un problème paramétrique d’écoulement autour d’un profil NACA0012. La prédiction pour un nouveau paramètre surpasse les modèles linéaires traditionnels lorsqu’elle est comparée aux simulations du modèle complet
Évaluation acoustique et aérodynamique du rotor DLR de petite échelle dans le cadre du GARTEUR AG 26
No full textInternational audienceThis paper presents the activities performed in the GARTEUR Action Group HC/AG-26 to study the acoustic and aerodynamic characteristics of small rotor configurations, including the influence of the rotor-rotor interactions. This paper will focus on comparisons between numerical activities and wind tunnel results on a small rotor provided by DLR. The wind tunnel models included a Rotor/Rotor/Pylon in isolated, tandem and coaxial configuration. The wind tunnel experiments for acoustics were performed in DLR’s Acoustic Wind Tunnel Braunschweig (AWB) and PIV test were performed in CIRA within a joint CIRA/DLR test program. For simulations, the numerical approaches from each partner are applied. The aerodynamic simulations necessary for the aeroacoustic predictions are conducted with various fidelity numerical methods, varying from lifting line to CFD. The acoustic values on the microphone positions are evaluated using Ffowcs Williams/Hawking (FW-H) formulation by all partners. The acoustic and aerodynamic predictions are compared to test data, including performance, PIV and acoustic directivity.Cet article présente les activités réalisées dans le cadre du GARTEUR HC/AG-26 dédié à l'étude des caractéristiques acoustiques et aérodynamiques de petites configurations de rotor, notamment l'influence des interactions rotor-rotor. Cet article se concentrera sur les comparaisons entre les activités numériques et les résultats en soufflerie sur un petit rotor fourni par le DLR. Des configurations isolée, tandem et coaxiale sont considérées. Les expériences en soufflerie pour l'acoustique ont été réalisées dans la soufflerie acoustique du DLR à Braunschweig (AWB) et les tests PIV ont été réalisés au CIRA dans le cadre d'un programme de test commun CIRA/DLR. Pour les simulations, les approches numériques de chaque partenaire sont appliquées. Les simulations aérodynamiques nécessaires aux prédictions aéroacoustiques sont réalisées avec une large gamme de méthodes numériques, allant de la ligne portante à la CFD. Les prédictions acoustiques et aérodynamiques sont comparées aux données de test, y compris les performances, la PIV et les directivités acoustiques
Évaluation des fermetures quadratiques pour la prédiction du décollement de coin sur une jonction aile-fuselage avec un angle de flèche négatif
No full textInternational audienceThe work proposed in this paper aims at improving the physical understanding and modelling of junction flows by conducting an analysis of several QCRs (Quadratic Constitutive Relations) applied to corner vortices. RANS computations are performed on two academic configurations for verification purposes and on a technical application of wing-body juncture featuring corner flows to achieve a better comprehension of the QCR variants ability to reproduce turbulent stress combinations relevant to secondary flows generation. This addresses the more theoretical interest of analysing the onset of corner separation (angle of attack 11 degree) and the role of corner flow vortices. Linear eddy viscosity model and the Extended QCR failed in detecting these vortical structures, the latter encountered numerical stability difficulties.On the other hand, QCR2000, QCR2020, QCR(r) showed a behaviour consistent with a simulation using Reynolds Stress Model, experiments and literature results. This better prediction was found to be related to the ability of these models to accurately detect the stress-induced vortex that delays and reduces the size of the corner separation. Notably, the more recent QCR(r) and QCR2020 have enhanced the estimation of the normal stress combination compared to the more widespread QCR2000, demonstrating their validity as alternative models.The work proposed in this paper aims at improving the physical understanding and modelling of junction flows by conducting an analysis of several QCRs (Quadratic Constitutive Relations) applied to corner vortices. RANS computations are performed on two academic configurations for verification purposes and on a technical application of wing-body juncture featuring corner flows to achieve a better comprehension of the QCR variants ability to reproduce turbulent stress combinations relevant to secondary flows generation. This addresses the more theoretical interest of analysing the onset of corner separation (angle of attack 11 deg) and the role of corner flow vortices. Linear eddy viscosity model and the Extended QCR failed in detecting these vortical structures, the latter encountered numerical stability difficulties. On the other hand, QCR2000, QCR2020, QCR(r) showed a behaviour consistent with a simulation using Reynolds Stress Model, experiments and literature results. This better prediction was found to be related to the ability of these models to accurately detect the stress-induced vortex that delays and reduces the size of the corner separation. Notably, the more recent QCR(r) and QCR2020 have enhanced the estimation of the normal stress combination compared to the more widespread QCR2000, demonstrating their validity as alternative models
Land Surface Temperature Super-Resolution with a Scale-Invariance-Free Neural Approach: Application to MODIS
No full textInternational audienceDue to the trade-off between the temporal and spatial resolution of thermal spaceborne sensors, super-resolution methods have been developed to provide fine-scale Land Surface Temperature (LST) maps. Most of them are trained at low resolution but applied at fine resolution, and so they require a scale-invariance hypothesis that is not always adapted. The main contribution of this work is the introduction of a Scale-Invariance-Free approach for training Neural Network (NN) models, and the implementation of two NN models, called Scale-Invariance-Free Convolutional Neural Network for Super-Resolution (SIF-CNN-SR) for the super-resolution of MODIS LST products. The Scale-Invariance-Free approach consists on training the models in order to provide LST maps at high spatial resolution that recover the initial LST when they are degraded at low resolution and that contain fine-scale textures informed by the high resolution NDVI. The second contribution of this work is the release of a test database with ASTER LST images concomitant with MODIS ones that can be used for evaluation of super-resolution algorithms. We compare the two proposed models, SIF-CNN-SR1 and SIF-CNN-SR2, with four state-of-the-art methods, Bicubic, DMS, ATPRK, Tsharp, and a CNN sharing the same architecture as SIF-CNN-SR but trained under the scale-invariance hypothesis. We show that SIF-CNN-SR1 outperforms the state-of-the-art methods and the other two CNN models as evaluated with LPIPS and Fourier space metrics focusing on the analysis of textures. These results and the available ASTER-MODIS database for evaluation are promising for future studies on super-resolution of LST
Robust estimation of particle image density up to high values through probabilistic modelling
No full textInternational audienceWe introduce a new method for estimating the seeding density of particle images, based on a statistical model for their pixel intensity. The model leads to very simple equations allowing for a very robust estimation of the seeding density, but also of parameters of the particle intensity distribution and of image noise. This model has been tested so far on different synthetic cases. A preliminary study has been made on a random distribution of points in the image plane spread by a Point Spread Function (PSF). This mock-up case has been used to study the statistical convergence of our model. A second study has been made on more realistic synthetic cases, resulting from the projection of a 3D particle cloud through a set of cameras, mimicking e.g. a 3D-PTV experimental setup. In both previous cases, we assumed both the particles intensities and the noise to be Gaussian-distributed. In order to better model real experimental intensities distributions, a third study has been made with the same setup but with a χ2 particle intensity distribution, to account for the Mie scattering behaviour, resulting in an intensity roughly proportional to the square of the particle radius, the latter obeying a Gaussian distribution. As the square of a Gaussian distribution is closely related to the χ2 law, this model represents a step closer to fully taking into account experimental images. Finally, a fourth study has been made on synthetic images generated for the currently running 2nd LPT Challenge (https://www.ispiv2025.org/2LPTDA.html) to try and match the announced image densities. We report here the performances on these different cases. For each case, two different situations are analyzed: one in which the noise parameters are known or calculated through other methods, and one in which they are not. For the latter case, two slightly different calculation methods will be used and compared. These synthetic cases give very promising results, with estimation errors on the seeding densities lower than 4% when the noise features are unknown, and lower than 2% when they can be estimated, up to very high densities (0.25 particles per pixel, ppp)
Assimilation des données du coefficient de pression pour la correction du modèle de turbulence à l'aide de l'inversion de champ pour le profil supercritique transsonique
No full textInternational audienceDespite the increasing prevalence of LES (Large Eddy Simulation) and DNS (Direct Numerical Simulation), these computational methods remain prohibitively expensive for everyday design applications in industrial settings. Consequently, Reynolds-averaged Navier–Stokes (RANS) simulations are often favored by industry due to their lower computational cost. However, RANS models are known to lack fidelity in certain flow configurations, particularly separated flows and flows with strong adverse pressure gradients. To address this limitation, turbulence models can be enhanced through a "field inversion" methodology coupled with "machine learning" techniques. In this study, performed using the DG Aghora code, we investigate the assimilation of pressure coefficient data to correct the Spalart-Allmaras turbulence model. The geometry used for this study is the transonic supercritical OAT15A wing profile. It is well-known for shock wave/turbulent boundary-layer interaction, a phenomenon called transonic buffet. Three assimilation parameters were evaluated and successfully corrected the turbulence model to accurately reproduce the pressure coefficient obtained from experimental data. Notably, despite assimilating only wall-pressure data, the Mach number field appears to be well-corrected across the domain. A difference between the three optimization parameters is observed in the skin friction coefficient, which is attributed to the physical implications of each optimization parameter
An Ultrawideband Linearly Polarized Unit-Cell Design for Ka-band SatCom Transmitarrays
No full textInternational audienceThis paper presents an ultrawideband linearly polarized (LP) unit cell (UC) for Ka-band transmitarray (TA) SatCom applications. The proposed UC is based on wideband magneto-electric (ME) dipole antennas. Simulation results of the UC demonstrate a reflection bandwidth (BW) S11 < -10 dB of 52% around a central frequency of 24 GHz, covering both the uplink and downlink frequency bands of the SatCom Ka-band. To evaluate its performance within a transmitarray, a LP TA comprising 18×18 elements is designed and simulated. Simulation results of the TA at 27.5 GHz indicate a realized gain of 23.6 dBi, with a crosspolarization discrimination (XPD) superior to 75 dB, and a low side lobe level (SLL) of 12 dB, making this UC a good candidate for SatCom applications