21549 research outputs found
Sort by
Observations et modélisation des extrêmes de précipitation aux échelles saisonnière à interannuelle au Sahel
Get PDFLes événements extrêmes de précipitations (EEP) ont un impact fort sur les populations, les biens et les activités économiques au Sahel. Les deux dernières décennies ont vu leur fréquence d’occurrence et leur intensité augmenter, et les projections climatiques tendent à montrer que cela va continuer. Cependant, de nombreuses zones d’ombre demeurent sur les mécanismes pilotant leurs propriétés et leur variabilité. Cette thèse se propose d’aborder certaines des questions scientifiques sous-jacentes : Quelles sont, qualitativement et quantitativement, ces propriétés ? Quelles sont les mécanismes gouvernant la variabilité des EEPs, notamment aux échelles interannuelles ? Un EEP peut prendre des formes variées. Ici, la définition adoptée est la suivante : tout événements pluvieux dont le cumul journalier sur une surface de 1◦x1◦ est supérieur ou égal au 99ème percentile de la distribution empirique de la pluie journalière estimée à l’échelle de cette surface.
Pour répondre à ces questions, cette thèse se décompose en deux volets. Le premier volet se focalise sur la documentation des EEPs sahéliens. Cet objectif requiert a priori l’utilisation d’observations de précipitation, ce qui représente malheureusement une difficulté majeure à l’échelle du Sahel (faible densité du réseau de pluviomètres, accessibilité souvent restreinte). Grâce à une collaboration avec l’agence météorologique du Burkina Faso, nous avons pu avoir accès à un réseau de pluviomètres particulièrement dense pour la région, et adapté à la documentation des EEPs sur ce pays pour la période récente (2001-2013). Ce jeu de données montre que les EEPs se produisent en phase avec le cycle annuel de la mousson africaine et plus fréquemment pendant les années humides. Par ailleurs, la variabilité de leur cumul est dominée par la variabilité de leur nombre plutôt qu’un changement dans leur intensité. Pour étendre ces résultats à l’échelle du Sahel, nous avons considéré la quasi-totalité des produits grillés de précipitation (24 produits mélangeant différentes sources d’observations et algorithmes d’estimation de la pluie). Ces produits ont d’abord été évalués sur la Burkina Faso à l’aide de nos données de référence. Leur plus grande difficulté est d’estimer quantitativement l’intensité des EEPs. Cette difficulté se manifeste sur l’ensemble du Sahel par une forte dispersion de ces produits sur le 99ème percentile de la pluie journalière. Malgré tout, les produits disponibles s’accordent sur le couplage entre précipitation totale et EEPs sur le Sahel, aux échelles saisonnière et interannuelle.
Le second volet propose d’utiliser la modélisation pour étudier l’impact de la variabilité interannuelle des températures de surface de la mer de l’Atlantique tropical sur les EEPs sahéliens, avec au coeur de cette stratégie le modèle atmosphérique global ARPEGE-Climat développé au CNRM. La dernière version disponible de ce modèle a tout d’abord été évaluée sur plusieurs facettes du climat ouest-africain. Son principal défaut est un manque important de précipitation sur le Sahel, limitant son utilisation pour notre objectif. Nous avons cependant pu mettre en oeuvre une approche statistique récemment développée dans la communauté française de modélisation du climat permettant de re-calibrer efficacement le modèle et de le rendre plus adapté à nos préoccupations. Plusieurs configurations ont été identifiées. Une de celle-ci est ensuite utilisée pour étudier la réponse du climat sahélien à des anomalies chaudes ou froides dans l’Atlantique tropical
Flat plate boundary layer accelerated by shock wave propagation
Get PDFThe flat plate transitional boundary layer response to the acceleration induced by the shock wave propagation is studied using large-eddy simulations. The steady boundary layer global behaviour is first investigated before focusing on the transient response of a turbulent region following the shock wave propagation. It is shown that the transient response of the turbulent region exhibits strong similarities with the spatial transition process to turbulence induced by free-stream turbulence, the so-called bypass transition. The boundary layer does not evolve gradually from the initial turbulence intensity to the final turbulence intensity but undergoes a temporal transition process composed of three distinct phases. These three different phases are comparable with the three stages of a bypass transition (i.e. buffeted laminar flow, transition and fully turbulent) because they are governed by the same physical processes. On the other hand, it is highlighted that this temporal response is identical to that described by He & Seddighi (J. Fluid Mech., vol. 715, 2013, pp. 60–102) during the study of an incompressible boundary layer undergoing an increase of mass flow rate. The boundary layer compression by the shock propagation does not contribute to any significant change in the turbulence dynamic after an unsteady acceleration
A mixed-categorical data-driven approach for prediction and optimization of hybrid discontinuous composites performance
Get PDFSurrogate models are an essential engineering tool and their popularity has increased recently due to the high computational cost of evaluating real-world simulations. However, most of these functions are described by mixed variables (continuous and categorical), which makes it harder to create accurate interpolation functions. This work builds a surrogate model from a given mixed data set, in order to quickly and accurately calculate the mechanical performance of hybrid discontinuous composites. Then, in order to find the optimal hybridization, three different approaches are performed: mono-objective, targeted and multi-objective. Starting from a virtual database provided by the industrial partner, the mixed categorical optimization process is performed by coupling a multi-armed bandit strategy with a continuous Bayesian optimization solver. The efficiency of the proposed approach is tested and two main results are achieved. The obtained surrogate models are shown to be sufficiently accurate, having an R² score grater than 90% in average. Our proposed optimization process is also able to identify correctly the optimal fibres with respect to the desirable targets
Learning-Enhanced Adaptive Robust GNSS Navigation in Challenging Environments
Get PDFGlobal Navigation Satellite System (GNSS) is the widely used technology when it comes to outdoor positioning. But it has severe limitations with regard to safety-critical applications involving unmanned autonomous systems. Namely, the positioning performance degrades in harsh propagation environment such as urban canyons. In this paper we propose a new algorithm for GNSS navigation in challenging environments based on robust statistics. M-estimators showed promising results in this context, but are limited by some fixed hyper-parameters. Our main idea is to adapt this parameter, for the Huber cost function, to the current environment in a data-driven manner. Doing so, we also present a simple yet efficient way of learning with satellite data, whose number may vary over time. Focusing the learning problem on a single parameter enables to efficiently learn with a lightweight neural network. The generalization capability and the positioning performance of the proposed method are evaluated in multiple contexts scenarios (open-sky, trees, urban and urban canyon), with two distinct GNSS receivers, and in an airplane ground inspection scenario. The maximum positioning error is reduced by up to 68% with respect to M-estimators
An approach for joint scheduling of production and predictive maintenance activities
Get PDFThe Industry 4.0 paradigm, thanks to the deployment of cutting-edge technologies enabling the deployment of new services, contributes to improve the agility of productive organizations. Among these services, the Prognostic and Health Management (PHM) contributes to the health assessment of the manufacturing resources and to prognose their future conditions by providing decision supports for production and predictive maintenance management. However, the future conditions of technical production resources depend on the productive tasks they will have to carry out. If their future conditions will not satisfy production criteria, maintenance tasks will have to be planned and productive tasks will be delayed or assigned to other resources for which their future conditions considering these new tasks must be assessed. In this context, a multi-agent system SCEMP (Supervisor, Customers, Environment, Maintainers and Producers) is here proposed in which production scheduling and predictive maintenance planning collaborate and exploit decision supports provided by PHM modules. The proposed multi-agent system provides a framework in which production and the predictive maintenance activities can be scheduled simultaneously by compromising on their objectives. During the scheduling process, SCEMP enables to identify the needed predictive maintenance from the assignments of production tasks to machines, the machine component prognoses and machine models. It schedules production tasks and predictive maintenance activities according to the number, competencies and availabilities of production and maintenance resources. The SCEMP framework is described and presented in the tough job shop context. For this context, case studies have been generated and scheduled within acceptable computation times. To illustrate the SCEMP functioning, some simplified case studies are detailed with the obtained performances. It is flexible and can be adapted to various manufacturing situations. It can also be used to assess the interest of implementing prognostic functions for machine components
Stability analysis of the JCAPL equivalent fluid model equations for porous media
Get PDFThe equivalent fluid model (EFM) describes the
acoustic properties of rigid porous media by defin-
ing it as a fluid with an effective density and
an effective compressibility. Their definition are
based on the dynamic tortuosity α and the dy-
namic compressibility β, known to be complex-
valued functions depending on frequency. Among
the different models describing α and β, this pa-
per focuses on the Johnson-Champoux-Allard-
Pride-Lafarge (JCAPL) model [1] where these
parameters are defined as irrational functions,
behaving like fractional derivatives in the time
domain. Here, we present the proof of stabil-
ity of the time-domain EFM using the JCAPL
model thanks to their oscillatory-diffusive (OD)
representations
PFEM: a mixed structure-preserving discretization method for port-Hamiltonian systems
Get PDFPFEM: a mixed structure-preserving discretization method for port-Hamiltonian systems
Excitation of instabilies in a Blasius boundary layer by surface vibration
Get PDFHere we have demonstrated that small amplitude vibration can artificially excite both
two dimensional (2D) and three dimensional (3D) instability modes. The 2D modes were
typical of Tollmien–Schlichting (TS) waves provided that the frequency of excitation lies
within the unstable region of the neutral stability predicted by modal linear stability
theory. However, even if the frequency of the mechanically forced mode was within the
stable bounds of the neutral curve the harmonics generated by the non-linear response
of the flow could develop as instability modes. Further analysis of the streamwise
and spanwise evolution of the instability modes identified from the temporal Fourier
transform confirmed the presence of 3D modes excited due to the nature of the mode
shape deflection of the vibrating panel which was not uniform in the spanwise direction.
The effect of spanwise non-uniformity could be increased by activating the motors
along the spanwise direction. However, due to the forcing from a combination of
both streamwise and spanwise motors, strong interaction with the 3D mode led to a
reduction in the growth rate of the TS wave in the far-field region despite higher initial
perturbation generated by a larger number of motors
Modeling, robust control synthesis and worst-case analysis for an on-orbit servicing mission with large flexible spacecraft
Get PDFThis paper outlines a complete methodology for modeling an on-orbit servicing mission scenario and
designing a feedback control system for the attitude dynamics that is guaranteed to robustly meet point-
ing requirements, despite model uncertainties as well as large inertia and flexibility changes throughout
the mission scenario. A model of the uncertain plant was derived, which fully captures the dynamics
and couplings between all subsystems as well as the decoupled/coupled configurations of the chaser/tar-
get system in a single linear fractional representation (LFR). In addition, a new approach is proposed to
model and analyze a closed-loop kinematic chain formed by the chaser and the target spacecraft through
the chaser’s robotic arm, which uses two local spring-damper systems with uncertain damping and stiff-
ness. This approach offers the possibility to model the dynamical behavior of a docking mechanism with
dynamic stiffness and damping. The controller was designed by taking into account all the interactions
between subsystems and uncertainties as well as the time-varying and coupled flexible dynamics. Lastly,
the robust stability and worst-case performances were assessed by means of a structured singular value
analysis. The main contribution of this paper is thus to fill an important gap in the literature by obtaining
a full analytical LFR model of a rendezvous on-orbit servicing mission including all the different phases
of such a scenario, namely the approach phase, capture/docking and manipulation of a target satellite,
while taking into account all parametric uncertainties and varying geometrical configurations
Influence of the fluid–fluid drag on the pressure drop in simulations of two-phase flows through porous flow cells
No full textTo macroscopically describe two-phase flows in porous media, we need accurate modeling of the drag forces between the two fluids and the solid phase. In low-permeability porous media, where capillarity is often dominant, momentum exchange is often neglected and the fluid–fluid drag force is treated as part of the drag between fluids and solid in the momentum transport equation. Two-phase flows in highly permeable porous media, however, are often characterized by a larger interface area between the two fluids and by thin films developing. In such cases, the fluid–fluid drag may play an important role and require a specific description in the momentum transport equations. Here, we use computational methods to study immiscible cocurrent two-phase flows in a microfluidic device made of an array of cylinders squeezed between two plates in a Hele-Shaw cell. The key idea is to solve 2D Stokes–Darcy equations integrated over the height of the cell, allowing us to explore different permeability ranges by changing the gap between the plates while keeping the in-plane 2D geometry in the cell unchanged. We use this approach to ask whether the fluid–fluid drag forces affect the pressure drop and how the permeability modifies the relative importance of the drag forces. We find different behaviors depending on the gap thickness, but the fluid–fluid drag plays a significant role in all cases