HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
Conception et entraînement de modèles de substitution pour l’estimation de la traînée automobile, à partir de champs géométriques et physiques avec un budget très limité de simulations haute-fidélité
No full textIn automotive engineering, optimizing vehicle shape is essential to reduce aerodynamic drag and improve energy efficiency. In the final stages of development, the challenge lies in refining the design to balance performance and aesthetics. Due to the high cost of CFD simulations, an exhaustive exploration of the design space is hardly feasible, highlighting the need for reduced-order models that enable fast drag estimation. This thesis proposes new methodologies to predict aerodynamic drag and physical fields from a limited number of CFD simulations. Two categories of models have been developed. Near field models predict physical quantities on the vehicle surface, such as pressure and viscous forces. This approach relies on the assumption that these quantities can be approximated using local geometric descriptors and physical variables from CFD data. Far field models estimate flow properties in the vehicle wake by leveraging a reparameterization of shape and reduced modes of physical fields to predict flow patterns on a downstream cross-sectional plane. The various models demonstrate strong predictive performance on geometries of industrial complexity and allow for fast evaluation of shape variation effects. These results show that, regardless of the chosen approach, combining local geometric information with physical data extracted from simulations enables accurate drag estimation, even with a limited dataset. These methods pave the way for fast and accurate predictive models, well-suited for low-cost interactive aerodynamic optimization.En ingénierie automobile, l’optimisation de la forme des véhicules est essentielle pour réduire la traînée aérodynamique et améliorer l’efficacité énergétique. En phase finale de développement, l’enjeu est d’affiner le design pour concilier performance et esthétique. En raison du coût élevé des simulations CFD, une exploration exhaustive de l’espace de conception n’est pas possible, d’où la nécessité de développer des modèles réduits permettant une estimation rapide de la traînée. Cette thèse propose de nouvelles méthodologies pour prédire la traînée aérodynamique et les champs physiques à partir d’un nombre limité de simulations CFD. Deux catégories de modèles ont été développées. Les modèles near field prédisent les grandeurs physiques à la surface du véhicule, comme la pression et les efforts visqueux. Cette approche repose sur l’hypothèse que ces quantités peuvent être approchées à partir de descripteurs géométriques locaux et de variables physiques issues des simulations CFD. Les modèles far field estiment les propriétés de l’écoulement dans le sillage du véhicule en exploitant un reparamétrage de la forme et des modes réduits des champs physiques pour prédire les écoulements sur un plan de coupe en aval. Les différents modèles montrent de très bonnes performances prédictives sur des géométries de complexité industrielle et permettent une évaluation rapide des effets des variations de forme. Ces travaux démontrent que, quelle que soit l’approche adoptée, la combinaison d’informations géométriques locales et de données physiques extraites des simulations permet une estimation précise de la traînée, même avec un jeu de données restreint. Ces méthodes ouvrent ainsi la voie à des modèles de prédiction rapides et précis, adaptés à l’optimisation aérodynamique interactive à coût réduit
P30-39 Inhalation Permeability Assessment Using Two Human Lung Epithelial Cell Models in Air- and Liquid-Liquid interface Cultures
No full textInternational audienceBackground & Purpose:Inhalation is a major route of chemical exposure in gaseous or particulate forms. However, the fate of inhaled chemicals and their ability to cross the pulmonary barrier remain poorly understood due to accessibility challenges and limitations of animal models. In vitro cell culture models are widely employed to assess chemical permeability, offering valuable insights into chemical absorption. However, these models often rely on nominal concentrations to determine the cell apparent permeability coefficients, which often results in inaccurate representations of actual cellular exposure due to chemical partitioning into various assay components. This study aimed to better understand the fate of four chemicals (propranolol, suma-triptan succinate, hippuric acid, phthalic acid mono 2-ethylhexyl) in permeability systems (Transwell® plates) and their transport across bronchial (Calu-3) and alveolar (h-AELVi) human lung epithelial cell models cultured under either air- (ALI) or liquid -liquid (LLI) interface conditions.Methods:First, this study compares bronchial and alveolar human lung epithelial cell models, regarding their capacity to form a tight epithelial cell barrier for a 2-week culture period in ALI and 1-week in LLI. Then, cytotoxicity assays were performed to evaluate the potential for damage to the cell barrier. After that, chemical evaporation and non-specific binding to polymers were evaluated in a cell-free in vitro permeability system. Furthermore, Rapid Equilibrium Dialysis (RED) was employed to determine non-specific binding to medium proteins and cellular proteins. Finally, the chemical transport was measured in apical to basolateral (A-B) for both cell models cultured in ALI or LLI.Results:Calu-3 and h-AELVi cells consistently form and maintain a tight epithelial barrier, regardless of the culture conditions. Two different no toxic concentrations (1 &10 µM) were used for the different assays. The chemical loss was observed in the permeability system (cell and plastic binding, volatilization, abiotic degradation, etc.). It varies depending on the chemical. Despite the observed differences in cell type or culture conditions, the permeability of the cells used was found to be nearly identical.Conclusion:Neglecting the chemical losses underestimates the real permeability. Therefore, accurately assessing lung model permeability requires considering chemical loss as opposed to reliance on nominal concentration. This approach is crucial for advancing physiologically based kinetic (PBK) models, as it ensures robust parameter derivation and significantly strengthens the predictive validity of human exposure and risk assessments
Études de soutenabilité utilisant des données provenant des systèmes d'information d'entreprise : Revue de littérature et feuille de route pour une recherche préliminaire.
No full textInternational audienceÉtudes de soutenabilité utilisant des données provenant des systèmes d'information d'entreprise : Revue de littérature et feuille de route pour une recherche préliminaire
Reliable Multi-Level Optimization for Safe Predictive Control of Autonomous Vehicles to Avoid Uncertain Multimodal PLEVs
No full textInternational audienceSafety assurance using all perceptual information to predict the motion of dynamic agents is critical in urban environments and remains an open challenge. For Autonomous Vehicles (AV) operating around vulnerable road users, the risk assessment strategy often needs to address stochastic uncertainties in the multiple possible trajectories (or multimodal motion) of the surrounding traffic agents. However, this increases the complexity of the navigation problem using the existing planners. To address this issue, this paper presents a multilevel optimization strategy that combines sampling-based and direct optimization methods for decision-making and control with improved safety and trajectory smoothness. In the primary stage, a sampling-based optimization framework systematically identifies safe candidate trajectories by employing the Fusion of stochastic Predictive Inter-Distance Profile (F-sPIDP). F-sPIDP encapsulates the multimodal dynamics of traffic agents and explicitly computes the uncertainties in their estimated or tracked states. From the set of trajectories, a reference optimal trajectory and its F-sPIDP setpoints are selected, adhering to stringent safety constraints and motion smoothness. Subsequently, a secondary local control optimization refines the optimal trajectory to ensure compliance with the AV's kinematic and dynamic constraints while accounting for the quantified uncertainty within the F-sPIDP framework. The performance of the proposed method was assessed through simulations and statistical analyses, evaluating its robustness to diverse levels of uncertainty.</div
Improving Vulnerable Road-Users Detection through Hybrid Collaborative Perception and Detection Refinement
No full textInternational audienceEnsuring the safety of autonomous vehicles in complex urban environments critically depends on accurate 3D object detection. While LiDAR sensors provide reliable depth information, their effectiveness is limited by sparsity at long distances and occlusions, particularly in intersection scenarios. Collaborative perception addresses these challenges by enabling information sharing among vehicles and infrastructure sensors, with intermediate fusion offering a balance between communication efficiency and detection accuracy. However, existing collaborative perception frameworks exhibit a notable performance gap between detecting vehicles and vulnerable road users such as cyclists and pedestrians. In this work, we propose a novel hybrid collaboration framework designed to reduce this gap. Our method leverages late-stage information from communicating agents to augment the ego agent's point cloud, then applies a standard intermediate fusion strategy, followed by a refinement stage that further improves the detection accuracy of various objects. Experiments on the Mixed Signals dataset demonstrate that our approach sets a new state-of-the-art in the detection of vulnerable road users in urban V2X scenario
Uncertainty Measures in a Generalized Theory of Evidence
No full textInternational audienceEpistemic Random Fuzzy Set theory is an extension of Dempster-Shafer and possibility theories in which pieces of evidence are represented by random fuzzy sets and combined by the productintersection rule, an extension of Dempster's rule and the product combination of possibility distributions. We propose a measure of imprecision and a measure of conflict for random fuzzy sets, uniquely characterized by minimal sets of requirements. Both measures have simple expressions involving only the contour function: in the finite case, imprecision is measured by the logarithm of the sum of the plausibility of the singletons, while conflict is measured by the negative logarithm of the maximum plausibility over the singletons. These definitions can be easily carried over to random fuzzy sets in continuous spaces, allowing us to define the imprecision and conflict of Gaussian random fuzzy numbers and extensions. Total uncertainty is defined as the sum of imprecision and conflict. The corresponding measure, referred to as \calT-entropy, happens to be the min-entropy and the nonspecificity measure of, respectively, the probability distribution and the normalized possibility distribution constructed from the contour function. The application of these uncertainty measures to belief elicitation is discussed and illustrated by some examples
Graded electrospun scaffold from aligned fibers to honeycomb micropatterns: Application to bone-tendon tissue engineering
No full textInternational audienceScaffolds' production for hard to soft tissues recently become of great interest, as bone-tendon insertion tissue engineering, where injuries mainly occur. Interfacial tissue engineering aims at developing grafts to mimic the gradients of those tissues as far as composition, mechanical properties and structures are concerned. Additive manufacturing can offer solutions to meet these requirements, but still requires to improve processes to achieve such gradients in a few steps.In this study, we developed a 3D-printed collector to combine gap-spinning and micropatterning. We were able to manufacture a scaffold (60 mm long, 5 mm wide) with a smooth gradient of 5 mm long from honeycomb structure to aligned fibers (promoting bone and tendon fate, respectively) in a single step. We estimated a gradient in Young modulus from 20 MPa to 30 MPa from the bone to the tendon side. Deformation tracking permitted to highlight significant difference of local strains between both areas, which could then impact cells' response. Murine stem cells C3H10T1/2 were then seeded at both scaffold parts and cultivated without any growth factors in stretching conditions. Alkaline phosphatase staining and tenomodulin immunostaining suggested the effect of stretching to cells' behavior between the bone and tendon area, compared to static condition. However, benefit of topographical and mechanical cues only cannot be fully established to foster cells to specific fate probably due to the limits of this cell line. This novel collector system however permitted to produce a relevant scaffold to study interfaces where a topographical gradient might be needed.</div
Angle-dependent acoustic performance of Green facades in the low frequency: effects of foliage thickness and density
No full textInternational audienceIn dense urban environments, road traffic noise is a significant health and environmental concern. While green facades are traditionally valued for their thermal and aesthetic benefits, they are increasingly being investigated for their acoustic potential. This study examines how the incidence angle of sound waves emitted by a road vehicle influences the acoustic reflection and insertion loss of green building facades in the low frequency range. Using a combination of the transfer matrix formalism and the Horoshenkov-Miki model, the present work analyzes the acoustic behavior of green facades across the 100–1000 Hz frequency range, considering variations in foliage thickness and leaf density. The resulting reflection coefficient mappings and sound pressure levels are computed for both individual frequencies and full spectrum conditions. The findings demonstrate that quarter-wave interference patterns, which depend on frequency and incidence angle, significantly affect acoustic performance. Insertion losses due to foliage range from 0.1 dB to 15 dB, underscoring the strong directional and spectral sensitivity of these systems. This variability largely overlooked in conventional acoustic models emphasizes the importance of accounting for directional sound propagation in urban noise simulations. The study offers insights for the acoustic optimization of green facades, contributing to more realistic noise control strategies aligned with public health policies and the lived experience of city resident
Digital twins for optimising intra-arterial therapy in liver cancer
No full textInternational audienceSelective internal radiation therapy (SIRT) has emerged as an effective and safe treatment for patients with unresectable and chemorefractory hepatocellular carcinoma (HCC). It consists of intra-arterial administration of radioactive microspheres, typically loaded with Yttrium-90, via a catheter directly into the hepatic artery. Unfortunately, sub-optimal efficacy of the SIRT can occur, and consequently, optimising the treatment is of utmost importance. In this work, we aim to systematically evaluate the role of the different factors that may influence the treatment success through a digital twin of the liver vasculature specific to the patient. The digital twin allowed us to reproduce the intra-arterial administration of Therasphere® in a patient diagnosed with two HCCs. An optimisation strategy was then applied to estimate the effect of the injection parameters and to identify the optimal injection configuration. The digital twin showed that the longitudinal position of the catheter within the arterial tree plays a major role in the targeting of the tumor. In addition, the catheter's tip inclination with respect to the hepatic arterial vessels was also proved to influence the microspheres' delivery to the tumor and the non-tumor tissuesThe developed digital twin and optimisation strategy could help the clinicians in personalising the SIRT, thus improving the targeting of the tumor(s) while preserving healthy tissues