e-book Portrait des musiciens à lheure du numérique (CEPREMAP) (French Edition)

Contents:

Citation data for document RePEc:cpm:opuscl

De plus, le modele de la plaque plane a ete utilise afin de realiser une etude parametrique portant sur le positionnement des actuateurs, l'effet de la rigidite ainsi que de l'epaisseur de la plaque. Finalement, la plaque plane a ete degivree en chambre climatique. Des cas de degivrage ont ete simules numeriquement afin d'etudier la possibilite d'utiliser un critere base sur la deformation pour predire le succes du systeme. La validation experimentale a confirme la capacite du logiciel a calculer precisement a la fois les frequences et les modes de resonance d'une structure et a simuler leur excitation par des actuateurs piezoelectriques.

L'etude revele que la definition de l'amortissement dans le modele numerique est essentiel pour l'obtention de resultats precis. Les resultats de l'etude parametrique ont demontre l'importance de minimiser l'epaisseur et la rigidite afin de reduire la valeur des frequences. Numerical simulation of two phase flows in heat exchangers; Simulation numerique des ecoulements diphasiques dans les echangeurs. The report presents globally the works done by the author in the thermohydraulic applied to nuclear reactors flows.

It presents the studies done to the numerical simulation of the two phase flows in the steam generators and a finite element method to compute these flows. Contributions to reinforced concrete structures numerical simulations ; Contributions a la simulation numerique de structures en beton arme. In order to be able to carry out simulations of reinforced concrete structures, it is necessary to know two aspects: This work deals with these two subjects.

After an accurate estimation of two behaviour models micro-plan and mesoscopic models , two damage models the first one using a scalar variable, the other one a tensorial damage of the 2 order are proposed. These two models belong to the framework of generalized standard materials, which renders their numerical integration easy and efficient.

A method of load control is developed in order to make easier the convergence of the calculations. At last, simulations of industrial structures illustrate the efficiency of the method. Direct numerical simulation of granular flows with fluid; Simulation numerique directe d'ecoulements granulaires en presence de fluide. Numerical models applied to simulation of granular flow with fluid are developed.

The physical model selected to describe particles flow is a discrete approach. Particle trajectories are calculated by the Newton law and collision is describe by a soft-sphere approach. The fluid flow is modelled by Navier-Stokes equations. The modelling of the momentum transfer depends on the resolution scale: The direct model is used to find representative elementary volume and prove the local character of the Ergun's law.

This application shows the numerical mesh size , physical Reynolds number and computational CPU time and memory consumptions limitations. The drag law model and the direct model are validated with analytical and empirical solutions and compared. For the two models, the CPU time and the memory consumptions are discussed.

The drag law model is applied to the simulation of gas-solid dense fluidized-beds. In the case of uniform gas distribution, the fluidized-bed simulation heights are compared to experimental data for particle of group A and B of the Geldart classification. Dry corrosion prediction of radioactive waste containers in long term interim storage: In the framework of research on long term behaviour of radioactive waste containers, this work consists on the one hand in the study of low temperature oxidation of iron and on the other hand in the development of a numerical model of oxide scale growth.

Isothermal oxidation experiments are performed on pure iron at and C in dry and humid air at atmospheric pressure. Oxide scales formed in these conditions are characterized. They are composed of a duplex magnetite scale under a thin hematite scale. The inner layer of the duplex scale is thinner than the outer one. Both are composed of columnar grains, that are smaller in the inner part.

The outer hematite layer is made of very small equiaxed grains. A model for iron oxide scale growth at low temperature is then deduced. It is based on the calculation of concentration profiles of chemical species and also point defects in the oxide scale and in the substrate. Indeed, these point defects can either be eliminated by interface motion or injected in the substrate, where they can be annihilated, considering sinks as the climb of dislocations.

Hence, the influence of substrate cold-work can be investigated. The EKINOX model is validated in the conditions of Wagner's theory and is confronted with experimental results by its application to the case of high temperature oxidation of nickel. Directory of Open Access Journals Sweden.

Observation des influences du changement du couvert forestier sur le comportement hydrologique de grands bassins versants tropicaux a l'aide de la teledetection numerique: Cas du bassin versant de Dong Nai, Viet Nam. Monitoring hydrological behavior of a large tropical watershed following a forest cover variation has an important role in water resource management planning as well as for forest sustainable management. Traditional methods in forest hydrology studies are Experimental watersheds, Upstream-downstream, Experimental plots, Statistical regional analysis and Watershed simulation.

Those methodes have limitations for large watersheds concerning the monitoring time, the lack of input data especially about forest cover and the capacity of extrapolating results accurately in terms of large watersheds. Moreover, there is still currently a scientific debate in forest ecology on relation between water and forest. The reason of this problem comes from geographical differences in publication concerning study zones, experimental watershed size and applied methods.

It gives differences in the conclusions on the influence of tropical forest cover change on the changes of outlet water and yet on the yearly runoff in terms of large watershed. In order to exceed the limitations of actual methods, to solve the difficulty of acquiring forest cover data and to have a better understanding of the relation between tropical forest cover change and hydrological behavior evolution of a large watershed, it is necessary to develop a new approach by using numeric remote sensing.

We used the watershed of Dong Nai as a case study. Flooding frequency and flooding peaks have clearly decreased when there is an increase of the forest cover from to The influence of tropical forest cover on the hydrological behavior is varying with geographical locations of watershed. There is a significant relation between forest cover evolution and environmental facteurs as the runoff. Experimental and numerical study of the active control of jets inside combustion chambers; Etude experimentale et numerique du controle actif de jets dans des chambres de combustion.

Combustion instabilities occur when the flame heat release couples with the acoustic waves propagating in the combustion chamber. This phenomenon can lead to strong vibrations and noise but also, sometimes, to the complete combustion device failure. That is the reason why so many studies focus on the control of those instabilities. The method chosen in this study consists in an active control device or set of actuators having a strong effect on the mixing of the burner exhaust flow with the ambient fluid. The model configuration studied consists in a non reactive jet of air controlled by four small tangential secondary jets.

Experiments have been carried out to optimize the control device geometry. The configuration identified as the most efficient, in terms of mixing enhancement, has been simulated through Large Eddy Simulations LES. The objective of the numerical part of the present work is double. First, the numerical simulations provide a better understanding of the phenomena occurring when the control is on. Then, it is shown that LES can be considered as a tool to predict the effects of a control device on a flow. Experimental study and numerical simulation of free pulsed jets; Etude experimentale et modelisation numerique des jets libres pulses.

A plane pulsed jet flow has been simulated by a finite difference method. Experimental results have also been obtained by laser tomography and particle image velocimetry. The results show that the flow is affected by the pulsation in the vicinity of the nozzle to reach an asymptotic state of a permanent jet. Comprendre le mouvement humain mobilise des chercheurs de nombreuses disciplines scientifiques: It allows reaching peak currents of hundred of amperes at average energies higher than 2 MeV.

The original concept of two accelerating cavities aims at minimizing the transverse and longitudinal emittances following the Gao's principles. From practical reasons the operating parameters, particularly the laser pulse duration, do not correspond to those considered in the design. Hence, numerical simulations were performed to evaluate the gun's performances in experimental environment.

Portrait d'Hélios Azoulay, compositeur et musicien

The study of a stabile injector operation resulted in evolutions with consequences in the phase control systems implying the laser and the HF Hyper Frequency source. The beam transverse and longitudinal characteristics have been measured as a function of the main parameters i. Measurements of the transverse emittance energy dispersion and wave packed duration are presented for several injector configurations.

The systems of existing beam measurements have been studied to determine the resolution and the experimental conditions to fulfill, in order to suggest improvements for the CANDELA beam. The experiments with the beam have been compared with numerical simulations. Agreement was obtained within wide ranges of parameters for most of the characteristic beam quantities. Based on a work at www. Asymptotic study and numerical simulation of laser wave propagation in an inhomogeneous medium; Etude asymptotique et simulation numerique de la propagation laser en milieu inhomogene.

To simulate the propagation of a monochromatic laser beam in a medium, we use the paraxial approximation of the Klein-Gordon in the time-varying problem and of the Maxwell in the non time-depending case equations. In a first part, we make an asymptotic analysis of the Klein-Gordon equation.

We obtain approximated problems, either of Schroedinger or of transport-Schroedinger type. We prove the existence and uniqueness of a solution for these problems, and estimate the difference between it and the exact solution of the Klein-Gordon equation. In a second part, we study the boundary problem for the advection Schroedinger equation, and show what the boundary condition must be so that the problem on our domain should be the restriction of the problem in the whole space: In a third part, we use the preceding results to build a numerical resolution method, for which we prove stability and show some simulations.

Modelling and numerical simulation of liquid-vapor phase transitions; Modelisation et simulation numerique des transitions de phase liquide-vapeur. This work deals with the modelling and numerical simulation of liquid-vapor phase transition phenomena. The study is divided into two part: In the first part, we study the classical viscous criteria for selecting weak solutions of the system used when the equation of state is non monotonic. Those criteria do not select physical solutions and therefore we focus a more recent criterion: We use this criterion to exactly solve the Riemann problem which imposes solving an algebraic scalar non linear equation.

Unfortunately, this step is quite costly in term of CPU which prevent from using this method as a ground for building Godunov solvers. That is why we propose an alternative approach two equations of state. Using the least action principle, we propose a phase changing two-phase flow model which is based on the second thermodynamic principle. We shall then describe two equilibrium submodels issued from the relaxations processes when instantaneous equilibrium is assumed.

Despite the weak hyperbolicity of the last sub-model, we propose stable numerical schemes based on a two-step strategy involving a convective step followed by a relaxation step. We show the ability of the system to simulate vapor bubbles nucleation. Large scale simulation numerical study of transition to turbulence in jets; Etude numerique par simulation des grandes echelles de la transition a la turbulence dans les jets. This study highlights the potentialities of the numerical technique of large scale simulation in describing and understanding the turbulent flows in a complex geometry.

The multiple coherent vortical structures implied in the transition to turbulence of the jet were found. At the same time, helical or annular axisymmetric vortices were observed. Also, an original vortical arrangement was evidenced, resulting from the alternating inclination and local pairing of these rings. It could been forced through an ad-hoc excitation which modifies subsequently drastically the jet development. When an axisymmetric excitation is imposed after formation of annular structures, pairs of counter-rotative longitudinal vortices occur and generate lateral jets.

Their nature and presence in case of a helical excitation are discussed. An efficient method for controlling their number is developed. Then, one is studied the very low frequency periodic phenomenon of backward-facing transition to turbulence which develops in the confined jet and grid multiple jets a phenomenon generic in numerous flows.

It was found to depend not only on the characteristic of the re-circulation pre-transition zones but also on the upstream flow zone of post-transition stagnation, pressure effect. Large scale transversal motions of the fluid have been found beginning from the grid. An interpretation of this phenomenon is suggested refs. Macroscopic numerical simulation model of multi-constituent fluid flows in porous medium; Modele macroscopique de simulation numerique d'ecoulements de fluides multiconstituants en milieu poreux.

This model should allow to foresee the quantitative composition of fluids in oil fields and also to improve the knowledge of the flow of different species inside mixtures. The overall physical phenomena taking place at oil fields is explained in the first chapter. Chapter 2 recalls some thermodynamical notions at the equilibrium and outside equilibrium. These notions, necessary to understand the forecasting methods used by petroleum geologists, are described in chapter 3.

This chapter includes also a bibliographic study about the methods of simulation of mass and heat transfers in porous media. In chapter 4, using the thermodynamical relations of irreversible processes described in chapter 2, a new type of macroscopic model allowing to describe the overall phenomena analyzed is developed.

The numerical method used to solve this new system of equations is precised. Finally, chapter 5 proposes a set of cases for the validation of the uncoupled phenomena and some qualitative examples of modeling of coupled phenomena. Derivation of the low Mach number diphasic system. Numerical simulation in mono-dimensional geometry; Derivation du systeme diphasique bas Mach.

Simulation numerique en geometrie monodimensionnelle. This work deals with the derivation of a diphasic low Mach number model obtained through a Mach number asymptotic expansion applied to the compressible diphasic Navier Stokes system, expansion which filters out the acoustic waves. This approach is inspired from the work of Andrew Majda giving the equations of low Mach number combustion for thin flame and for perfect gases.

When the equations of state verify some thermodynamic hypothesis, we show that the low Mach number diphasic system predicts in a good way the dilatation or the compression of a bubble and has equilibrium convergence properties. Then, we propose an entropic and convergent Lagrangian scheme in mono-dimensional geometry when the fluids are perfect gases and we propose a first approach in Eulerian variables where the interface between the two fluids is captured with a level set technique.

In view of understanding and forecasting pollutant dispersion in urban areas, high resolution numerical simulations are performed. The aim is to reproduce atmospheric characteristics above complex urbanised site. An accurate method is developed to implement numerical simulations of the urban atmosphere based on three complementary tools, optimized on Marseille agglomeration example: Eulerian numerical simulation of gas-solid flows with several particles species; Modelisation numerique eulerienne des ecoulements gaz-solide avec plusieurs especes de particules.

The simulation of the multiphase flows is currently an important scientific, industrial and economic challenge. The objective of this work is to improve comprehension via simulations of poly-dispersed flows and contribute the modeling and characterizing of its hydrodynamics. The study of gas-solid systems involves the models that takes account the influence of the particles and the effects of the collisions in the context of the momentum transfer. This kind of study is covered on the framework of this thesis. Simulations achieved with the Saturne-polyphasique-Tlse code, developed by Electricite de France and co-worked with the Institut de Mecanique des Fluides de Toulouse, allowed to confirm the feasibility of approach CFD for the hydrodynamic study of the injectors and dense fluidized beds.

The stages of validation concern, on the one hand, the placement of the tool for simulation in its current state to make studies of validation and sensitivity of the models and to compare the numerical results with the experimental data. In addition, the development of new physical models and their establishments in the code Saturne will allow the optimization of the industrial process.

To carry out this validation in a satisfactory way, a key simulation is made, in particular a monodisperse injection and the radial force of injection in the case of a poly-disperse flow, as well as the fluidization of a column made up of solid particles. In this last case, one approached three configurations of dense fluidized beds, in order to study the influence of the grid on simulations ; then, one simulates the operation of a dense fluidized bed with which one characterizes the segregation between two various species of particles.

The study of the injection of the poly-disperse flows presents two configurations; a flow Co-current gas-particle in gas Case Hishida , and in addition, a poly-phase flow in a configuration of the jet type confined with zones of recirculation and stagnation case. Micromechanical simulation of Uranium dioxide polycrystalline aggregate behaviour under irradiation; Modele numerique micro-mecanique d'agregat polycristallin pour le comportement des combustibles oxydes.

A multi scale and multi-physic approaches are needed for the simulation of fuel behavior under irradiation. The main phenomena to take into account are thermomechanical behavior of the fuel rod and chemical-physic behavior of the fission products. These last years one of the scientific issue to improve the simulation is to take into account the multi-physic coupling problem at the microscopic scale.

The objective of this ph-D study is to contribute to this multi-scale approach. Mean field and full field approaches are considered. For the former and the later a self consistent homogenization technique and a periodic Finite Element model base on the 3D Voronoi pattern are respectively used. A cohesive zone model has also been developed and implemented to simulate grain boundary sliding and intergranular crack opening. The effective homogenous behaviour of a Representative Volume Element RVE is fitted with experimental data coming from mechanical tests on a single pellet.

Local behavior is also analyzed in order to evaluate the model capacity to assess micro-mechanical state. In particular, intra and inter granular stress gradient are discussed. A first validation of the local behavior assessment is proposed through the simulation of intergranular crack opening measured in a compressive creep test of a single fuel pellet. Concerning the impact of the microstructure on the fuel behavior under irradiation, a RVE simulation with a representative transient loading of a fuel rod during a power ramp test is achieved.

The impact of local stress and strain heterogeneities on the multi. Numerical modelling of fire propagation: Electricite de France, wishing to limit the accidental unavailability of its nuclear plants and to ensure their safety rigorously takes particular care to reduce the risk of fire. In this context, the Heat Transfer and Aerodynamics Branch of the Research and Development Division has been in charge of the design of numerical tools to simulate the fire propagation in buildings since Its program is articulated towards three axes which include: This paper gives on overview of the activity in progress in this research fields.

It illustrates also the applications performed and anticipated at Electricite de France of the numerical simulation in fire safety design. We discuss at the end of it the limitations and the development factors of these tool use. Mechanical characterization and numerical modeling of a rotary X-ray anode; Caracterisation mecanique et simulation numerique d'une anode tournante de rayons X.

This works deals with the design of light rotary anodes used for the generation of X-rays in medical scanners. Such anodes are made of graphite coated with tungsten by low pressure plasma sputtering. The mechanical behaviour of these materials during intense thermo-mechanical solicitation has been studied.

In a first step, the in-service conditions of solicitation are defined in terms of excitation frequency, temperature, deformation and deformation velocity. The analysis of used anodes has permitted to define the main modes of in-service damage. Tests were performed on small size samples over the complete temperature range between ambient temperature and deg. Carbon has shown a fragile elastic behaviour while tungsten has shown a more complex behaviour: C, then plastic, and becoming creep sensible above deg. Original load paths have permitted to show the existence of an internal back-stress and a coupling between plastic and viscous deformations.

The definition of an original phenomenological law of behaviour with a double inelastic, plastic and visco-plastic deformation and with an interaction term between both flow mechanisms has been necessary to describe the mechanical behaviour of tungsten.

Driftwood;
!
Home - Professor Marc Bourreau!
Welche Gründe führten zum Bau der Mauer in Berlin? Und welche Gründe führten zu deren Ende? (German Edition).
FIVE SIMPLE WAYS TO SAVE COSTS IN YOUR BUSINESS TODAY.
Are You an Author?!

An axisymmetrical 2-D mesh has permitted to calculate the constraints generated by the post-annealing cooling, by one and several series of radiographies and finally by a complete cooling after use. The repetition of radiographies rapidly leads to stabilized cycles. The calculated stress levels are realistic and. Numerical modelling 2 D and 3 D of circulating fluidized bed: The numerical modelling of internal CFB boilers flows faced with complex phenomenons due to the flows un-stationariness, the heterogeneousness of the particle size distribution, and interactions between the two phases and the walls.

Our study consisted in applying numerical models to the experimental configuration of cold circulating fluidized bed studied at the Cerchar. Then, to evaluate the validity limits of this model, we have built the regime diagram, and we have compared it with the experimental diagram. We have also noticed the importance of the choice of the mean diameter of the simulated particles. Elaboration of an alpha-numeric classification for file of matters of the documentation service of the CEA; Elaboration d'une classification alfha- numerique pour le fichier matieres du service de documentation du Commissariat a l'Energie Atomique.

Centre d' Etudes Nucleaires.

We give the principles of a classification of matters to square basis, suiting the needs of the Service, of Documentation of the C. We present the detail of the categories in the order of the 'columns', likewise the big scientific subdivisions at the CEA. On presente ensuite le detail des rubriques dans l'ordre des ''colonnes'', c'est-a-dire, des grandes subdivisions scientifiques du C. Contributions to a rational design of heterogeneous catalysts: I present through this dissertation a synthesis of my contributions to the field of heterogeneous catalysis, along two decades of research undertaken as a scientist at Institut Francais du Petrole.

The presentation of my work follows therefore a guideline starting with preparation and ending at modelization of the catalytic solid, after having gone through its characterization and the assessment of its activity. Modelization is thus founded on a consistent set of experimental informations. This guideline is applied to the four main themes to which this work is confined: In summary, I believe I have contributed significantly, on the one hand to strong conceptual and technical advances in the area of ab initio simulation of elementary phenomena in heterogeneous catalysis, with the elaboration of original knowledge on catalysis by sulfides, metals and acids, as well as the genesis of alumina carriers, and on the other hand to a new approach of periodic trends in catalysis: In a near future it will be possible to say if practical results validate this conceptual tool, and justify or not the ambitious title I gave to my work.

Publications in International Journals

Numerical modelling in building thermo-aeraulics: For 3D modelling of thermo-aeraulics in building using field codes, it is necessary to reduce the computing time in order to model increasingly larger volumes. The solution suggested in this study is to couple two modelling: The first part of the work that was carried out is the setting of a simplified CFD modelling. We propose rules for use of coarse grids, a constant effective viscosity law and adapted coefficients for heat exchange in the framework of building thermo-aeraulics.

The second part of this work concerns the creation of fluid Macro-Elements and their coupling with a calculation of CFD finite volume type. Depending on the boundary conditions of the problem, a local description of the driving flow is proposed via the installation and use of semi-empirical evolution laws. We use these two approaches on five cases representative of thermo-aeraulics in buildings. The results are compared with experimental data and with traditional RANS simulations. We highlight the significant gain of time that our approach allows while preserving a good quality of numerical results.

Numerical and experimental study of the mixture of engine jets in the wake vortices of an airline aircraft; Etude numerique et experimentale du melange des jets de moteur dans les tourbillons de sillage d'un avion de ligne. This study is a contribution to the understanding of the formation and duration of aircraft condensation trails. The development of a numerical code based on the direct resolution of the 3-D compressible Navier-Stokes equations has been done first. Then, an experiment has been carried out in a wind tunnel to analyze the problem of the mixture of heated jets in a wing wake.

A first validation of the numerical method has been carried out from bibliographic results and measurements of the mixture evolution of an inert tracer contained in the engine jets during a flight test. In order to characterize the condensation inside the wake, the evolution of the local water vapor saturation ratio has been calculated. The influence of the Crow instability on the mixture of effluents in the high atmosphere is also shown. Finally, a comparison is made between the numerical simulation results and the experimental measurements obtained in this study.

The numerical results have also permitted to characterize the low scale exchange mechanisms between a turbulent jet and a swirl flow. Simulation du potentiel energetique de la houle au port autonome Climate change in cities due to global warming For that we carried out the microscopic simulations , based on the Monte. Eastwood, "Computer simulation using particles",. Developpement D'un Modele Climatique Regional: Le sujet de cette these concerne la modelisation numerique du climat regional.

L'objectif principal de l'exercice est de developper un modele climatique regional ayant les capacites de simuler des phenomenes de meso-echelle spatiale. Notre domaine d'etude se situe sur la Cote Ouest nord americaine. Ce dernier a retenu notre attention a cause de la complexite du relief et de son controle sur le climat. Les raisons qui motivent cette etude sont multiples: Jusqu'alors, les MCG constituaient les modeles les plus estimes pour leurs aptitudes a simuler le climat ainsi que les changements climatiques mondiaux.

Toutefois, les phenomenes climatiques de fine echelle echappent encore aux MCG a cause de leur faible resolution spatiale. De plus, les repercussions socio-economiques des modifications possibles des climats sont etroitement liees a des phenomenes imperceptibles par les MCG actuels. Afin de circonvenir certains problemes inherents a la resolution, une approche pratique vise a prendre un domaine spatial limite d'un MCG et a y imbriquer un autre modele numerique possedant, lui, un maillage de haute resolution spatiale.

Ce processus d'imbrication implique alors une nouvelle simulation numerique. Cette "retro- simulation " est guidee dans le domaine restreint a partir de pieces d'informations fournies par le MCG et forcee par des mecanismes pris en charge uniquement par le modele imbrique. Ainsi, afin de raffiner la precision spatiale des previsions climatiques de grande echelle, nous developpons ici un modele numerique appele FIZR, permettant d'obtenir de l'information climatique regionale valide a la fine echelle spatiale.

Explicit dynamics for numerical simulation of crack propagation by the extended finite element method; Dynamique explicite pour la simulation numerique de propagation de fissure par la methode des elements finis etendus. Computerized simulation is nowadays an integrating part of design and validation processes of mechanical structures. Simulation tools are more and more performing allowing a very acute description of the phenomena. Moreover, these tools are not limited to linear mechanics but are developed to describe more difficult behaviours as for instance structures damage which interests the safety domain.

A dynamic or static load can thus lead to a damage, a crack and then a rupture of the structure. The fast dynamics allows to simulate 'fast' phenomena such as explosions, shocks and impacts on structure. The application domain is various. It concerns for instance the study of the lifetime and the accidents scenario of the nuclear reactor vessel. It is then very interesting, for fast dynamics codes, to be able to anticipate in a robust and stable way such phenomena: The extended finite element method has the advantage to break away from mesh generation and from fields projection during the crack propagation.

Effectively, crack is described kinematically by an appropriate strategy of enrichment of supplementary freedom degrees. Difficulties connecting the spatial discretization of this method with the temporal discretization of an explicit calculation scheme has then been revealed; these difficulties are the diagonal writing of the mass matrix and the associated stability time step. Here are presented two methods of mass matrix diagonalization based on the kinetic energy conservation, and studies of critical time steps for various enriched finite elements. The interest revealed here is that the time step is not more penalizing than those of the standard finite elements problem.

Comparisons with numerical simulations on another code allow to validate the theoretical works. A crack propagation test in mixed mode has been exploited in order to verify the simulation. Nine aspects were discussed, giving a state of the art in the domain: Numerical simulation of the unsteady and turbulent flow in a high-pressure turbine stage; Simulation numerique de l'ecoulement instationnaire et turbulent dans un etage de turbine haute pression. The aim of this study concerns the use of numerical methods for the resolution of the Reynolds Averaged Navier Stokes equations adapted to the simulation of the cooling of the trailing edge of a stator in a high pressure turbine.

The scheme is applied through a finite volume approach on control volume centered on the cells of a multi-block structured mesh. A simulation of the flow in a bidimensional stator, without cooling, is carried out. The cooling, which is realized with trailing edge slots, is then simulated on a bidimensional stator. Because the slot is represented by meshes overlapping the mesh of the smooth blade, the Chimera method is chosen.

This method makes it possible computations with overlapping meshes. The comparison with the experimental data, on these two first computations has validated this strategy to represent such slots. The tridimensional simulation of a single stator with taking account of the cooling is then realized. It showed the complex and tridimensional aspects of the main flow with focus on the influence of the cooling system. Finally two steady computations, without and with cooling, and an unsteady computation without cooling are carried out on a high pressure turbine stage.

The comparison with the experimental data obtained in the frame of the European Brite-Euram program is made. These results make it possible to determine the effect of the cooling on the flow in a turbine stage. Parallel direct numerical simulation of turbulent flows in rotor-stator cavities. Turbulent flows between a fixed disc and a rotating disc are encountered in various applications like turbo-machineries or torque converters of automatic gear boxes.

The results presented are restricted to the comparison between results obtained with direct simulation and results obtained with the MATHILDA code in the same configuration. Evaluation of turbulent transport and flame surface dissipation using direct numerical simulation of turbulent combustion; Evaluation des termes de transport et de dissipation de surface de flamme par simulation numerique directe de la combustion turbulente. The assumption of gradient transport for the mean reaction progress variable has a limited domain of validity in premixed turbulent combustion. The existence of two turbulent transport regimes, gradient and counter-gradient, is demonstrated in the present work using Direct Numerical Simulations DNS of plane flame configurations.

The DNS data base describes the influence of the heat release factor, of the turbulence-to-flame velocity ratio, and of an external pressure gradient. The simulations reveal a strong correlation between the regime of turbulent transport and the turbulent flame speed and turbulent flame thickness. These effects re not well described by current turbulent combustion models. Furthermore, he development of flame instabilities in turbulent configurations is also observed in the simulations.

A criterion is derived that determines the domain of occurrence of these instabilities Darrieus- Landau instabilities, Rayleigh- Taylor instabilities, thermo-diffusive instabilities. This criterion suggests that the domain of occurrence of flame instabilities is not limited to small Reynolds numbers. Premixed turbulent flame-wall interaction is studied using theoretical and numerical analysis. Laminar interactions are first investigated through a literature review.

This gives a characterization of the different configurations of interaction and justifies the use of simplified kinetic schemes to study the interaction. Calculations are then performed using Direct Numerical Simulation with a one-step chemistry model, and are compared with good agreements to asymptotic analysis. Flame-wall distances and wall heat fluxes obtained are compared successfully with those of the literature. Heat losses decrease the consumption rate, leading to extinction at the maximum of wall heat flux. It is followed by a flame retreat, when the fuel diffuses into the reaction zone, resulting in low unburnt hydrocarbon levels.

Then, turbulent regime is investigated, using two types of Direct Numerical Simulations: Similar results are obtained: In the 2D simulations , minimal flame-wall distances and maximum wall heat fluxes are similar to laminar values. However, the structure of the turbulence in the 3D calculations induces smaller flame-wall distances and higher wall heat fluxes.

Finally, a flame-wall interaction model is built and validated. It uses the flamelet approach, where the flame is described in terms of consumption speed and flame surface density. It is validated in an engine calculation. Modelling and numerical simulation of two-phase flows using the two-fluid two-pressure approach; Modelisation et simulation numerique des ecoulements diphasiques par une approche bifluide a deux pressions. This thesis is devoted to the modelling and numerical simulation of liquid-vapor flows. In order to describe these phase transition flows, a two-fluid two-pressure approach is considered.

This description of the liquid-vapor mixing is associated to the seven-equation model introduced by Baer and Nunziato. This work investigates the properties of this model in order to simulate the phase transition flows occurring in nuclear engineering. First, a theoretical thermodynamic framework is constructed to describe the liquid-vapor mixing. Provided with this framework, various modelling choices are suggested for the interaction terms between the phases.

These closure laws comply with an entropy inequality. The mathematical properties of this model are thereafter examined. The convective part is associated to a nonconservative hyperbolic system. First, we focus on the definition of its weak solutions. Several flow regimes for the two-phase mixing derive from this analysis. Such regimes for the two-phase flows are analogous to the torrential and fluvial regimes for the shallow-water equations.

Furthermore, we establish the linear and nonlinear stabilities of the liquid-vapor equilibrium. Finally, the implementation of a turbulence model and the introduction of a reconstruction process for the interfacial area are investigated in order to refine the description of the interfacial transfers. Using a fractional step approach, a Finite Volume method is at last constructed to simulate this model. First, various nonconservative adaptations of standard Riemann solvers are developed to approach the convective part.

Unlike the classic nonconservative framework, these schemes converge towards the same solution. Furthermore, a new relaxation scheme is proposed to approach the interfacial transfers. Provided with these schemes, the whole numerical method preserves the liquid-vapor equilibria. Using this numerical method, a careful comparison between the one- and two-pressure two.

A novel extrusion head for the renovation of cable sheaths - numeric simulation speeds conceptual work; Une tete d'extrusion novatrice pour revetir des cables. La simulation numerique pour accelerer sa conception. This article describes a new method of applying layers of insulation to power cables. The disadvantages of traditional methods of extrusion of cable insulation are looked at, that often are a source of impurities that can possibly degrade the physical and chemical qualities of the end product. The article discusses how the new method was developed in co-operation with a specialised University of Applied Science.

Also, the savings in time and money that can be made during the optimisation of a prototype are discussed.

Customers Also Bought Items By

The numeric simulation methods used are examined and the commercialisation of the product is looked at. Numerical simulations of flows through fixed networks of monodispersed and bi-dispersed spheres, for moderate Reynolds numbers; Simulations numeriques d'ecoulements a travers des reseaux fixes de spheres monodisperses et bidisperses, pour des nombres de Reynolds moderes. The application of statistically averaged two-fluid models for the simulation of complex indus- trial two-phase flows requires the development of adequate models for the drag force exerted on the inclusions and the interfacial heat exchange.

This task becomes problematic at high volume fractions of the dispersed phase. The quality of the simulation strongly depends upon the inter- facial exchange terms, starting with the steady drag force. For example, an accurate modelling of the drag force is therefore a crucial point to simulate the expansion of dense fluidized beds. Most models used to study the exchange terms between particles and fluids are based on the interaction between an isolated particle and a surrounding gas.

Those models are clearly not adequate in cases where the volume fraction of particles increases and particle-particle interactions become important. Studying such cases is a complex task because of the multiple possible configurations. While the interaction between an isolated sphere and a gas depends only on the particle size and the slip velocity between gas and particles, the interaction between a cloud of particles and a gas depends on many more parameters: Even if the particles keep relative fixed positions, there is an infinite number of combinations to construct such an array.

The objective of the present work is to perform steady and unsteady simulations of the flow in regular arrays of fixed particles in order to analyze the influence of the size and distributions of spheres on drag force and heat transfer the array of spheres can be either monodispersed, either bi-dispersed. Moreover some discrepancies are observed between the different studies. On top of that, all existing studies are limited to steady flows.

Direct numerical simulation and modeling of turbulent natural convection in a vertical differentially heated slot; Simulation numerique directe et modelisation de la convection naturelle turbulente dans un canal differentiellement chauffe. The main objectives of this thesis are the direct numerical simulation of natural convection in a vertical differentially heated slot and the improvements of second-order turbulence modelling.

A three-dimensional direct numerical simulation code has been developed in order to gain a better understanding of turbulence properties in natural convection flows. This code has been validated in several physical configurations: A detailed analysis of these results has revealed the principal qualities of the available models but has also pointed our their shortcomings.

This data base has been used in order to improve the triple correlations transport models and to select the turbulent time scales suitable for such flows. Tokamak electron heat transport by direct numerical simulation of small scale turbulence; Transport de chaleur electronique dans un tokamak par simulation numerique directe d'une turbulence de petite echelle. In a fusion machine, understanding plasma turbulence, which causes a degradation of the measured energy confinement time, would constitute a major progress in this field.

In tokamaks, the measured ion and electron thermal conductivities are of comparable magnitude. The possible sources of turbulence are the temperature and density gradients occurring in a fusion plasma. Whereas the heat losses in the ion channel are reasonably well understood, the origin of the electron losses is more uncertain. In addition to the radial velocity associated to the fluctuations of the electric field, electrons are more affected than ions by the magnetic field fluctuations. In experiments, the confinement time can be conveniently expressed in terms of dimensionless parameters.

The model is built by using the set of Braginskii equations. The developed simulation code is global in the sense that a fixed heat flux is imposed at the inner boundary, leaving the gradients free to evolve. From the nonlinear simulations , we have put in light three characteristics for the ETG turbulence: The thermal transport dependence study has shown a very small role of the normalized pressure, which is in contradiction with the Ohkama's formula.

On the other hand. Calculation of reactivity by digital processing; Calcul de la reactivite par traitement numerique. With a view to exploring the new possibilities offered by digital techniques, a description is given of the optimum theoretical conditions of a computation of the realtime reactivity using counting samples obtained from a nuclear reactor.

The degree to which these optimum conditions can be attained depends on the complexity of the processing which can be accepted. A compromise thus has to be made between the accuracy required and the simplicity of the equipment carrying out the processing. An example is given, using a relatively simple structure, which gives an idea of the accuracy of the results obtained over a wide range of reactor power. Ces conditions optimales peuvent etre approchees d'autant mieux que l'on accepte un traitement plus complexe.

Un compromis est donc a faire entre la precision desiree et la simplicite du materiel assurant le traitement. Un exemple adoptant une structure de complexite reduite permet de juger de la precision des resultats obtenus sur une importante plage d'evolution de la puissance. Towards a quantification of stress corrosion mechanisms: We discuss the respective roles played by anodic dissolution and hydrogen in SCC mechanisms of f. It is surprising to note that even the crystallographies at the scale of the micron are different, the macroscopic inclination of the fracture surfaces are the same.

In the case of L steel, the formation of strong pile-ups in the presence of hydrogen leads to a zigzag fracture along alternated slip planes in the most general case. In the absence of hydrogen, as in copper, this mechanism effectively disappears. Furthermore, numerical simulations of crack shielding by dislocations emitted on one plane predict the macroscopic inclination. It shows that it is due to the mere dissolution which confines slip activity at the very crack tip in f. In order to quantify the mechanism involved in L steel, we developed simulations which numerically solve the coupled diffusion and elasticity equations for hydrogen in the presence of a crack and shielding dislocations.

They reproduce the mechanisms of hydrogen segregation on edge dislocations and of a localised softening effect by decreasing pair interactions. These mechanisms lead to i a localisation of hydrogen embrittlement along the activated slip planes, ii an increase of the dislocation density in pile-ups, and iii a decrease of the cross slip probability. These three factors enhance micro-fracture at the head of a pile-up, which is responsible of thezigzag fracture. Introducing the free surface effects for hydrogen, we point out a new mechanism: The quantification of these different mechanisms allows to give a relation between the local fracture possibility and the macroscopic parameters.

A general law for.

Geological Survey USGS , in cooperation with the Du Page County Stormwater Management Department, maintains a database of hourly meteorological and hydrologic data for use in a near real-time streamflow simulation system. Generic tools for sensitivity analysis and code coupling are developed in the Caml language.

The user of these generic platforms has only to provide the specific part of the application in any language of his choice. Global sensitivity analysis of thermomechanical models in modelling of welding; Analyse de sensibilite globale de modeles thermomecanique de simulation numerique du soudage. Current approach of most welding modellers is to content themselves with available material data, and to chose a mechanical model that seems to be appropriate.

Among inputs, those controlling the material properties are one of the key problems of welding simulation: This way to proceed neglect the influence of the uncertainty of input data on the result given by the computer code. In this case, how to assess the credibility of prediction? This thesis represents a step in the direction of implementing an innovative approach in welding simulation in order to bring answers to this question, with an illustration on some concretes welding cases.

The global sensitivity analysis is chosen to determine which material properties are the most sensitive in a numerical welding simulation and in which range of temperature. Using this methodology require some developments to sample and explore the input space covering welding of different steel materials. Finally, input data have been divided in two groups according to their influence on the output of the model residual stress or distortion.

In this work, complete methodology of the global sensitivity analysis has been successfully applied to welding simulation and lead to reduce the input space to the only important variables. Sensitivity analysis has provided answers to what can be considered as one of the probable frequently asked questions regarding welding simulation: Global sensitivity analysis of thermo-mechanical models in numerical weld modelling; Analyse de sensibilite globale de modeles thermomecaniques de simulation numerique du soudage.

Contribution to the study of the behaviour of solid particles in a confined turbulent flow using direct numerical simulation ; Contribution a l'etude du comportement de particules solides en ecoulement turbulent confine par simulation numerique directe. The theme of this numerical thesis is on the behavior of solid particles embedded in a non-homogeneous and non-isotropic turbulent gas flow as the one tacking place in a plane channel.

The turbulence is generated through the direct numerical simulation of Navier-Stokes equations discretized by formally second order in time and space finite difference operators. This Eulerian vision of the incompressible gas flow is completed by a Lagrangian formulation allowing to follow solid particles. In this formulation, the considered forces are the non-linear drag and the Saffman lift both corrected for wall effects. Furthermore, depending on the test cases studied, particle bouncing forces on the wall, gravity or electrostatic forces are taken into account.

A three-dimensional Hermitian interpolation highlight the special care spend on the determination of the fluid velocity at the solid particle location. The first code application is dedicated to solid particles dispersion inside an horizontal channel, or in a channel operated in a weightlessness state. The huge amount of data from the Lagrangian tracking is reduced to the integral times of the turbulence seen by the solid particles on their trajectories.

Those times are crucial in Lagrangian methods associated with a low numerical cost compared with the ones used in the present study. Among those methods, the one based on Langevin type equations have the best potential to handle industrial type problems. Nevertheless, this method needs to rebuild the fluid velocity fluctuations seen by the solid particles on their trajectories.

This technic is able to reproduce the crossing trajectory effect, the inertial effect and the continuity effect, only if the integral times of the turbulence seen are known. Till now, those times were known thanks to a semi-empirical correlation from direct numerical simulation in homogeneous and isotropic turbulence Wang and Stock However, although these conditions, this correlation was. Numerical simulation and analysis of axial instabilities occurrence and development in turbomachines.

Application to a break transient in a helium nuclear reactor; Simulation numerique et analyse du declenchement et du developpement des instabilites axiales dans les turbomachines: The subject of the present work was to develop models able to simulate axial instabilities occurrence and development in multistage turbomachines. The construction of a 1D unsteady axisymmetric model of internal flow in a turbomachine at the scale of the row has followed different steps: This model is integrated in a numerical tool, which has the capacity to describe the gas dynamics in a complete circuit containing different elements ducts, valves, plenums.

Thus, the complete model can represent the coupling between local and global phenomena, which is a very important mechanism in axial instability occurrence and development. An elementary theory has also been developed, based on a generalisation of Greitzer's model. These models, which were validated on various configurations, have provided complementary elements for the validation of the complete model. They have also allowed a more comprehensive description of physical phenomena at stake in instability occurrence and development by quantifying various effects inertia, compressibility, performance levels and underlying the main phenomena in particular the collapse and recovery kinetics of the plenum , which were the only retained in the final elementary theory.

The models were first applied to academic configurations compression system , and then to an innovative industrial project: The use of the models have brought comprehensive elements to surge occurrence due to a break event. It has been shown that surge occurrence is highly dependent of break location and that surge. Metrological characterization of the numerical system Adonis for gamma spectrometry; Caracterisation metrologique du systeme de spectrometrie gamma numerique Adonis. In gamma spectrometry, new acquisition systems based on digital processing of the signals are now available on the market.

These tests have clearly shown that the performances announced by the manufacturers were generally not met. Feasibility of using of the second gradient theory for the direct numerical simulation of liquid-vapor flows with phase-change; Etude des potentialites de la theorie du second gradient pour la simulation numerique directe des ecoulements liquide-vapeur avec changement de phase.

One on the main difficulties encountered in the direct numerical simulation of two-phase flows in general and of liquid-vapor flows with phase-change in particular, is the interface tracking. The idea developed in this work consists in modeling a liquid-vapor interface as a volumetric zone across which physical properties vary continuously instead of a discontinuous surface. The second gradient theory allows to establish the evolution equations of the fluid in the whole system: That means that the resolution of a unique system of partial differential equations is necessary to determine the whole two-phase flow, the interfaces and their evolution in time being a part of the solution of this unique system.

We show in this work that it is possible to artificially enlarge an interface without changing its surface tension and the latent heat of vaporization. That means than it is possible to track all the interfaces of a liquid-vapor two-phase flow with phase-change on a mesh the size of which is imposed by the smallest Kolmogorov scale of the bulk phases for example.

The artificial enlargement of an interfacial zone is obtained by modifying the thermodynamic behavior of the fluid within the binodal. We show that this modification does not change the dynamics of an interface. However, although the thickness of an interface and its surface tension vary with the mass and heat fluxes that go though it, the thermodynamic modification necessary to the artificial enlargement of an interface drastically increases these variations.

Consequently, the artificial enlargement of an interface must be made carefully to avoid a too much important variation of its surface tension during dynamic situations. Contribution to the development of numerical tools for the hardening of electronic devices to the neutronic and electromagnetic environment induced by a high power gain shot; Contribution a la mise en oeuvre de moyens de simulation numerique pour l'etude de la vulnerabilite des systemes electriques soumis a l'environnement radiatif et electromagnetique du Laser Megajoule.

When fusion ignition will be attained inside the target chambers of high energy laser facilities LMJ-France and NIF-Usa , a harsh environment, composed of nuclear particles and an electromagnetic pulse EMP will be induced. All electronic devices located in the vicinity will be sensitive to this environment.

In the first part of this work, a simulation method has been developed to evaluate transient currents that will be induced in coaxial cables. The relevance of this model is then discussed thanks to comparisons with experimental results. In a second part, the possibility to simulate the propagation of the EMP, inside and outside such a big structure as a target chamber, using the finite difference in time and domain FDTD method is evaluated. The use of a classic FDTD method is impossible for this kind of simulation because of the huge computer resources needed.

It is the reason why a 3-dimensional space-time sub-grid method for FDTD has been developed and some massively parallel FDTD calculations have also been performed. Simulation numerique de l'accretion de glace sur une pale d'eolienne. The wind energy industry is growing steadily, and an excellent place for the construction of wind farms is northern Quebec. This region has huge wind energy production potential, as the cold temperatures increase air density and with it the available wind energy.

However, some issues associated with arctic climates cause production losses on wind farms. Icing conditions occur frequently, as high air humidity and freezing temperatures cause ice to build up on the blades, resulting in wind turbines operating suboptimally. One of the negative consequences of ice accretion is degradation of the blade's aerodynamics, in the form of a decrease in lift and an increase in drag. Also, the ice grows unevenly, which unbalances the blades and induces vibration. This reduces the expected life of some of the turbine components.

If the ice accretion continues, the ice can reach a mass that endangers the wind turbine structure, and operation must be suspended in order to prevent mechanical failure. To evaluate the impact of ice on the profits of wind farms, it is important to understand how ice builds up and how much it can affect blade aerodynamics. In response, researchers in the wind energy field have attempted to simulate ice accretion on airfoils in refrigerated wind tunnels.

Unfortunately, this is an expensive endeavor, and researchers' budgets are limited. However, ice accretion can be simulated more cost-effectively and with fewer limitations on airfoil size and air speed using numerical methods. Numerical simulation is an approach that can help researchers acquire knowledge in the field of wind energy more quickly. For years, the aviation industry has invested time and money developing computer codes to simulate ice accretion on aircraft wings.

Nearly all these codes are restricted to use by aircraft developers, and so they are not accessible to researchers in the wind engineering field. Moreover, these codes have been developed to meet aeronautical industry. Modelisation numerique et validation experimentale d'un systeme de protection contre le givre par elements piezoelectriques. Le degivrage au moyen d'actuateurs piezoelectriques est considere comme une avenue prometteuse pour le developpement de systemes a faible consommation d'energie applicables aux helicopteres legers.

Ce type de systeme excite des frequences de resonances d'une structure pour produire des deformations suffisantes pour rompre l'adherence de la glace. Par contre, la conception de tel systeme demeure generalement mal comprise. Ce projet de maitrise etudie l'utilisation de methodes numeriques pour assister la conception des systemes de protection contre le givre a base d'elements piezoelectriques. La methodologie retenue pour ce projet a ete de modeliser differentes structures simples et de simuler l'excitation harmonique des frequences de resonance au moyen d'actuateurs piezoelectriques.

Le calcul des frequences de resonances ainsi que la simulation de leur excitation a ensuite ete validee a l'aide de montages experimentaux. La procedure a ete realisee pour une poutre en porte-a-faux et pour une plaque plane a l'aide du logiciel de calcul par elements finis, Abaqus. De plus, le modele de la plaque plane a ete utilise afin de realiser une etude parametrique portant sur le positionnement des actuateurs, l'effet de la rigidite ainsi que de l'epaisseur de la plaque. Finalement, la plaque plane a ete degivree en chambre climatique. Des cas de degivrage ont ete simules numeriquement afin d'etudier la possibilite d'utiliser un critere base sur la deformation pour predire le succes du systeme.

La validation experimentale a confirme la capacite du logiciel a calculer precisement a la fois les frequences et les modes de resonance d'une structure et a simuler leur excitation par des actuateurs piezoelectriques. L'etude revele que la definition de l'amortissement dans le modele numerique est essentiel pour l'obtention de resultats precis.

Les resultats de l'etude parametrique ont demontre l'importance de minimiser l'epaisseur et la rigidite afin de reduire la valeur des frequences. Numerical simulation of two phase flows in heat exchangers; Simulation numerique des ecoulements diphasiques dans les echangeurs. The report presents globally the works done by the author in the thermohydraulic applied to nuclear reactors flows. It presents the studies done to the numerical simulation of the two phase flows in the steam generators and a finite element method to compute these flows.

Contributions to reinforced concrete structures numerical simulations ; Contributions a la simulation numerique de structures en beton arme. In order to be able to carry out simulations of reinforced concrete structures, it is necessary to know two aspects: This work deals with these two subjects. After an accurate estimation of two behaviour models micro-plan and mesoscopic models , two damage models the first one using a scalar variable, the other one a tensorial damage of the 2 order are proposed.

Citation data for document RePEc:cpm:opuscl

These two models belong to the framework of generalized standard materials, which renders their numerical integration easy and efficient. A method of load control is developed in order to make easier the convergence of the calculations. At last, simulations of industrial structures illustrate the efficiency of the method. Direct numerical simulation of granular flows with fluid; Simulation numerique directe d'ecoulements granulaires en presence de fluide. Numerical models applied to simulation of granular flow with fluid are developed. The physical model selected to describe particles flow is a discrete approach.

Particle trajectories are calculated by the Newton law and collision is describe by a soft-sphere approach. The fluid flow is modelled by Navier-Stokes equations. The modelling of the momentum transfer depends on the resolution scale: The direct model is used to find representative elementary volume and prove the local character of the Ergun's law. This application shows the numerical mesh size , physical Reynolds number and computational CPU time and memory consumptions limitations. The drag law model and the direct model are validated with analytical and empirical solutions and compared.

For the two models, the CPU time and the memory consumptions are discussed. The drag law model is applied to the simulation of gas-solid dense fluidized-beds. In the case of uniform gas distribution, the fluidized-bed simulation heights are compared to experimental data for particle of group A and B of the Geldart classification. Dry corrosion prediction of radioactive waste containers in long term interim storage: In the framework of research on long term behaviour of radioactive waste containers, this work consists on the one hand in the study of low temperature oxidation of iron and on the other hand in the development of a numerical model of oxide scale growth.

Both are composed of columnar grains, that are smaller in the inner part. The outer hematite layer is made of very small equiaxed grains. A model for iron oxide scale growth at low temperature is then deduced. It is based on the calculation of concentration profiles of chemical species and also point defects in the oxide scale and in the substrate.

Indeed, these point defects can either be eliminated by interface motion or injected in the substrate, where they can be annihilated, considering sinks as the climb of dislocations. Hence, the influence of substrate cold-work can be investigated.

François Moreau

The EKINOX model is validated in the conditions of Wagner's theory and is confronted with experimental results by its application to the case of high temperature oxidation of nickel. Directory of Open Access Journals Sweden. Observation des influences du changement du couvert forestier sur le comportement hydrologique de grands bassins versants tropicaux a l'aide de la teledetection numerique: Cas du bassin versant de Dong Nai, Viet Nam.

Monitoring hydrological behavior of a large tropical watershed following a forest cover variation has an important role in water resource management planning as well as for forest sustainable management. Traditional methods in forest hydrology studies are Experimental watersheds, Upstream-downstream, Experimental plots, Statistical regional analysis and Watershed simulation. Those methodes have limitations for large watersheds concerning the monitoring time, the lack of input data especially about forest cover and the capacity of extrapolating results accurately in terms of large watersheds.

Moreover, there is still currently a scientific debate in forest ecology on relation between water and forest. The reason of this problem comes from geographical differences in publication concerning study zones, experimental watershed size and applied methods. It gives differences in the conclusions on the influence of tropical forest cover change on the changes of outlet water and yet on the yearly runoff in terms of large watershed.

In order to exceed the limitations of actual methods, to solve the difficulty of acquiring forest cover data and to have a better understanding of the relation between tropical forest cover change and hydrological behavior evolution of a large watershed, it is necessary to develop a new approach by using numeric remote sensing.

The model configuration studied consists in a non reactive jet of air controlled by four small tangential secondary jets. Experiments have been carried out to optimize the control device geometry. The configuration identified as the most efficient, in terms of mixing enhancement, has been simulated through Large Eddy Simulations LES. The objective of the numerical part of the present work is double. First, the numerical simulations provide a better understanding of the phenomena occurring when the control is on.

Then, it is shown that LES can be considered as a tool to predict the effects of a control device on a flow. Experimental study and numerical simulation of free pulsed jets; Etude experimentale et modelisation numerique des jets libres pulses. A plane pulsed jet flow has been simulated by a finite difference method. Experimental results have also been obtained by laser tomography and particle image velocimetry. The results show that the flow is affected by the pulsation in the vicinity of the nozzle to reach an asymptotic state of a permanent jet.

Comprendre le mouvement humain mobilise des chercheurs de nombreuses disciplines scientifiques: It allows reaching peak currents of hundred of amperes at average energies higher than 2 MeV. The original concept of two accelerating cavities aims at minimizing the transverse and longitudinal emittances following the Gao's principles. From practical reasons the operating parameters, particularly the laser pulse duration, do not correspond to those considered in the design.

Hence, numerical simulations were performed to evaluate the gun's performances in experimental environment. The study of a stabile injector operation resulted in evolutions with consequences in the phase control systems implying the laser and the HF Hyper Frequency source. The beam transverse and longitudinal characteristics have been measured as a function of the main parameters i.

Measurements of the transverse emittance energy dispersion and wave packed duration are presented for several injector configurations. The systems of existing beam measurements have been studied to determine the resolution and the experimental conditions to fulfill, in order to suggest improvements for the CANDELA beam.

The experiments with the beam have been compared with numerical simulations. Agreement was obtained within wide ranges of parameters for most of the characteristic beam quantities. Based on a work at www. Asymptotic study and numerical simulation of laser wave propagation in an inhomogeneous medium; Etude asymptotique et simulation numerique de la propagation laser en milieu inhomogene. To simulate the propagation of a monochromatic laser beam in a medium, we use the paraxial approximation of the Klein-Gordon in the time-varying problem and of the Maxwell in the non time-depending case equations.

In a first part, we make an asymptotic analysis of the Klein-Gordon equation. We obtain approximated problems, either of Schroedinger or of transport-Schroedinger type. We prove the existence and uniqueness of a solution for these problems, and estimate the difference between it and the exact solution of the Klein-Gordon equation. In a second part, we study the boundary problem for the advection Schroedinger equation, and show what the boundary condition must be so that the problem on our domain should be the restriction of the problem in the whole space: In a third part, we use the preceding results to build a numerical resolution method, for which we prove stability and show some simulations.

Those criteria do not select physical solutions and therefore we focus a more recent criterion: We use this criterion to exactly solve the Riemann problem which imposes solving an algebraic scalar non linear equation. Unfortunately, this step is quite costly in term of CPU which prevent from using this method as a ground for building Godunov solvers.

That is why we propose an alternative approach two equations of state. Using the least action principle, we propose a phase changing two-phase flow model which is based on the second thermodynamic principle. We shall then describe two equilibrium submodels issued from the relaxations processes when instantaneous equilibrium is assumed.

Particularly, it is focussed on flows of free jet, confined jets and multiple jets of high solidity grid. Spatial simulations of the circular zone close to a free jet, of high Reynolds number were performed. In spite of an evident sensitivity to upstream conditions good agreement between our statistical predictions and different experimental measurements was obtained. The multiple coherent vortical structures implied in the transition to turbulence of the jet were found. At the same time, helical or annular axisymmetric vortices were observed. Also, an original vortical arrangement was evidenced, resulting from the alternating inclination and local pairing of these rings.

It could been forced through an ad-hoc excitation which modifies subsequently drastically the jet development. When an axisymmetric excitation is imposed after formation of annular structures, pairs of counter-rotative longitudinal vortices occur and generate lateral jets. Their nature and presence in case of a helical excitation are discussed. An efficient method for controlling their number is developed.

Then, one is studied the very low frequency periodic phenomenon of backward-facing transition to turbulence which develops in the confined jet and grid multiple jets a phenomenon generic in numerous flows. It was found to depend not only on the characteristic of the re-circulation pre-transition zones but also on the upstream flow zone of post-transition stagnation, pressure effect.

Large scale transversal motions of the fluid have been found beginning from the grid. An interpretation of this phenomenon is suggested refs. Macroscopic numerical simulation model of multi-constituent fluid flows in porous medium; Modele macroscopique de simulation numerique d'ecoulements de fluides multiconstituants en milieu poreux.

In chapter 4, using the thermodynamical relations of irreversible processes described in chapter 2, a new type of macroscopic model allowing to describe the overall phenomena analyzed is developed. The numerical method used to solve this new system of equations is precised.

Finally, chapter 5 proposes a set of cases for the validation of the uncoupled phenomena and some qualitative examples of modeling of coupled phenomena. Derivation of the low Mach number diphasic system. Numerical simulation in mono-dimensional geometry; Derivation du systeme diphasique bas Mach. Simulation numerique en geometrie monodimensionnelle. This work deals with the derivation of a diphasic low Mach number model obtained through a Mach number asymptotic expansion applied to the compressible diphasic Navier Stokes system, expansion which filters out the acoustic waves. This approach is inspired from the work of Andrew Majda giving the equations of low Mach number combustion for thin flame and for perfect gases.

Maya Bacache-Beauvallet

Simulations achieved with the Saturne-polyphasique-Tlse code, developed by Electricite de France and co-worked with the Institut de Mecanique des Fluides de Toulouse, allowed to confirm the feasibility of approach CFD for the hydrodynamic study of the injectors and dense fluidized beds. The stages of validation concern, on the one hand, the placement of the tool for simulation in its current state to make studies of validation and sensitivity of the models and to compare the numerical results with the experimental data.

In addition, the development of new physical models and their establishments in the code Saturne will allow the optimization of the industrial process. To carry out this validation in a satisfactory way, a key simulation is made, in particular a monodisperse injection and the radial force of injection in the case of a poly-disperse flow, as well as the fluidization of a column made up of solid particles. In this last case, one approached three configurations of dense fluidized beds, in order to study the influence of the grid on simulations ; then, one simulates the operation of a dense fluidized bed with which one characterizes the segregation between two various species of particles.

The main phenomena to take into account are thermomechanical behavior of the fuel rod and chemical-physic behavior of the fission products. These last years one of the scientific issue to improve the simulation is to take into account the multi-physic coupling problem at the microscopic scale. The objective of this ph-D study is to contribute to this multi-scale approach.

Mean field and full field approaches are considered. For the former and the later a self consistent homogenization technique and a periodic Finite Element model base on the 3D Voronoi pattern are respectively used. A cohesive zone model has also been developed and implemented to simulate grain boundary sliding and intergranular crack opening. The effective homogenous behaviour of a Representative Volume Element RVE is fitted with experimental data coming from mechanical tests on a single pellet.

Such anodes are made of graphite coated with tungsten by low pressure plasma sputtering. The mechanical behaviour of these materials during intense thermo-mechanical solicitation has been studied. In a first step, the in-service conditions of solicitation are defined in terms of excitation frequency, temperature, deformation and deformation velocity. The analysis of used anodes has permitted to define the main modes of in-service damage. Tests were performed on small size samples over the complete temperature range between ambient temperature and deg. Carbon has shown a fragile elastic behaviour while tungsten has shown a more complex behaviour: C, then plastic, and becoming creep sensible above deg.

Original load paths have permitted to show the existence of an internal back-stress and a coupling between plastic and viscous deformations. The definition of an original phenomenological law of behaviour with a double inelastic, plastic and visco-plastic deformation and with an interaction term between both flow mechanisms has been necessary to describe the mechanical behaviour of tungsten.

The cold-drawing generated by each flow is translated into kinematic variables. The numerical identification of the parameters has been performed using an optimizer coupled to a finite element code which simulates the flexural test. The obtained law has been validated by the experimental observation of paths for complex loads. This behaviour law has been finally used to simulate the conditions of use of a real anode. An axisymmetrical 2-D mesh has permitted to calculate the constraints generated by the post-annealing cooling, by one and several series of radiographies and finally by a complete cooling after use.

The repetition of radiographies rapidly leads to stabilized cycles. The calculated stress levels are realistic and. Numerical modelling 2 D and 3 D of circulating fluidized bed: The numerical modelling of internal CFB boilers flows faced with complex phenomenons due to the flows un-stationariness, the heterogeneousness of the particle size distribution, and interactions between the two phases and the walls.

Elaboration of an alpha-numeric classification for file of matters of the documentation service of the CEA; Elaboration d'une classification alfha- numerique pour le fichier matieres du service de documentation du Commissariat a l'Energie Atomique. Centre d' Etudes Nucleaires. We give the principles of a classification of matters to square basis, suiting the needs of the Service, of Documentation of the C.

We present the detail of the categories in the order of the 'columns', likewise the big scientific subdivisions at the CEA. On presente ensuite le detail des rubriques dans l'ordre des ''colonnes'', c'est-a-dire, des grandes subdivisions scientifiques du C. Contributions to a rational design of heterogeneous catalysts: I present through this dissertation a synthesis of my contributions to the field of heterogeneous catalysis, along two decades of research undertaken as a scientist at Institut Francais du Petrole.

I started my itinerary on the 'floor', with the task of developing industrial hydro-treating catalysts, then I had the nice opportunity to lead advanced research on various subjects. However, I have been devoting myself for the past ten years to the encounter between catalysis and theoretical chemistry. The presentation of my work follows therefore a guideline starting with preparation and ending at modelization of the catalytic solid, after having gone through its characterization and the assessment of its activity. Modelization is thus founded on a consistent set of experimental informations.

This guideline is applied to the four main themes to which this work is confined: In summary, I believe I have contributed significantly, on the one hand to strong conceptual and technical advances in the area of ab initio simulation of elementary phenomena in heterogeneous catalysis, with the elaboration of original knowledge on catalysis by sulfides, metals and acids, as well as the genesis of alumina carriers, and on the other hand to a new approach of periodic trends in catalysis: In a near future it will be possible to say if practical results validate this conceptual tool, and justify or not the ambitious title I gave to my work.

The solution suggested in this study is to couple two modelling: The first part of the work that was carried out is the setting of a simplified CFD modelling. We propose rules for use of coarse grids, a constant effective viscosity law and adapted coefficients for heat exchange in the framework of building thermo-aeraulics.

The second part of this work concerns the creation of fluid Macro-Elements and their coupling with a calculation of CFD finite volume type.

Portrait des musiciens à l'heure du numérique (CEPREMAP t. 22) (French Edition).
Publications in French Journals.
Anima immortale (Italian Edition)?
simulation numerique du: Topics by www.farmersmarketmusic.com!
;

Depending on the boundary conditions of the problem, a local description of the driving flow is proposed via the installation and use of semi-empirical evolution laws. We use these two approaches on five cases representative of thermo-aeraulics in buildings. The results are compared with experimental data and with traditional RANS simulations. We highlight the significant gain of time that our approach allows while preserving a good quality of numerical results.

Eastwood, "Computer simulation using particles",. Developpement D'un Modele Climatique Regional: Le sujet de cette these concerne la modelisation numerique du climat regional. L'objectif principal de l'exercice est de developper un modele climatique regional ayant les capacites de simuler des phenomenes de meso-echelle spatiale. Notre domaine d'etude se situe sur la Cote Ouest nord americaine. Ce dernier a retenu notre attention a cause de la complexite du relief et de son controle sur le climat.

Les raisons qui motivent cette etude sont multiples: Jusqu'alors, les MCG constituaient les modeles les plus estimes pour leurs aptitudes a simuler le climat ainsi que les changements climatiques mondiaux. Toutefois, les phenomenes climatiques de fine echelle echappent encore aux MCG a cause de leur faible resolution spatiale. De plus, les repercussions socio-economiques des modifications possibles des climats sont etroitement liees a des phenomenes imperceptibles par les MCG actuels.

Afin de circonvenir certains problemes inherents a la resolution, une approche pratique vise a prendre un domaine spatial limite d'un MCG et a y imbriquer un autre modele numerique possedant, lui, un maillage de haute resolution spatiale. Ce processus d'imbrication implique alors une nouvelle simulation numerique. Cette "retro- simulation " est guidee dans le domaine restreint a partir de pieces d'informations fournies par le MCG et forcee par des mecanismes pris en charge uniquement par le modele imbrique. Ainsi, afin de raffiner la precision spatiale des previsions climatiques de grande echelle, nous developpons ici un modele numerique appele FIZR, permettant d'obtenir de l'information climatique regionale valide a la fine echelle spatiale.

Simulation tools are more and more performing allowing a very acute description of the phenomena. Moreover, these tools are not limited to linear mechanics but are developed to describe more difficult behaviours as for instance structures damage which interests the safety domain. A dynamic or static load can thus lead to a damage, a crack and then a rupture of the structure. The fast dynamics allows to simulate 'fast' phenomena such as explosions, shocks and impacts on structure. The application domain is various.

It concerns for instance the study of the lifetime and the accidents scenario of the nuclear reactor vessel. It is then very interesting, for fast dynamics codes, to be able to anticipate in a robust and stable way such phenomena: The extended finite element method has the advantage to break away from mesh generation and from fields projection during the crack propagation. Effectively, crack is described kinematically by an appropriate strategy of enrichment of supplementary freedom degrees.

Difficulties connecting the spatial discretization of this method with the temporal discretization of an explicit calculation scheme has then been revealed; these difficulties are the diagonal writing of the mass matrix and the associated stability time step. Here are presented two methods of mass matrix diagonalization based on the kinetic energy conservation, and studies of critical time steps for various enriched finite elements.

The interest revealed here is that the time step is not more penalizing than those of the standard finite elements problem. Comparisons with numerical simulations on another code allow to validate the theoretical works. A crack propagation test in mixed mode has been exploited in order to verify the simulation.

Nine aspects were discussed, giving a state of the art in the domain: Numerical simulation of the unsteady and turbulent flow in a high-pressure turbine stage; Simulation numerique de l'ecoulement instationnaire et turbulent dans un etage de turbine haute pression.

The aim of this study concerns the use of numerical methods for the resolution of the Reynolds Averaged Navier Stokes equations adapted to the simulation of the cooling of the trailing edge of a stator in a high pressure turbine. The scheme is applied through a finite volume approach on control volume centered on the cells of a multi-block structured mesh. A simulation of the flow in a bidimensional stator, without cooling, is carried out. The cooling, which is realized with trailing edge slots, is then simulated on a bidimensional stator.

Because the slot is represented by meshes overlapping the mesh of the smooth blade, the Chimera method is chosen. This method makes it possible computations with overlapping meshes. The comparison with the experimental data, on these two first computations has validated this strategy to represent such slots.

The tridimensional simulation of a single stator with taking account of the cooling is then realized. It showed the complex and tridimensional aspects of the main flow with focus on the influence of the cooling system. Finally two steady computations, without and with cooling, and an unsteady computation without cooling are carried out on a high pressure turbine stage.

Evaluation of turbulent transport and flame surface dissipation using direct numerical simulation of turbulent combustion; Evaluation des termes de transport et de dissipation de surface de flamme par simulation numerique directe de la combustion turbulente. The assumption of gradient transport for the mean reaction progress variable has a limited domain of validity in premixed turbulent combustion. The existence of two turbulent transport regimes, gradient and counter-gradient, is demonstrated in the present work using Direct Numerical Simulations DNS of plane flame configurations.

Furthermore, he development of flame instabilities in turbulent configurations is also observed in the simulations. A criterion is derived that determines the domain of occurrence of these instabilities Darrieus- Landau instabilities, Rayleigh- Taylor instabilities, thermo-diffusive instabilities.

This criterion suggests that the domain of occurrence of flame instabilities is not limited to small Reynolds numbers. Premixed turbulent flame-wall interaction is studied using theoretical and numerical analysis. Laminar interactions are first investigated through a literature review. This gives a characterization of the different configurations of interaction and justifies the use of simplified kinetic schemes to study the interaction. Calculations are then performed using Direct Numerical Simulation with a one-step chemistry model, and are compared with good agreements to asymptotic analysis.

Flame-wall distances and wall heat fluxes obtained are compared successfully with those of the literature. Heat losses decrease the consumption rate, leading to extinction at the maximum of wall heat flux. It is followed by a flame retreat, when the fuel diffuses into the reaction zone, resulting in low unburnt hydrocarbon levels. Then, turbulent regime is investigated, using two types of Direct Numerical Simulations: