La présente thèse porte sur la modélisation dédiée au contrôle des phénomènes physiques se produisant au travers des restrictions de la chaîne d’air du moteur. L’objectif est d’assurer la précision des modèles tout en limitant les temps de calcul associés. En effet, les modèles dédiés au contrôle nécessitent un temps de calcul court (proche du temps réel) afin d’être exploitables dans un processus de développement. En parallèle de cela, les normes concernant les rejets polluants sont de plus en plus drastiques, et nécessitent de prendre en compte finement les phénomènes physiques mis en jeu. Dans une première partie, les équations des phénomènes physiques liés aux turbocompresseurs radiaux sont introduites dans une approche cartographique. Les équations fondamentales des turbomachines sont développées et utilisées afin de construire une méthode d’interpolation / extrapolation de champs turbocompresseur expérimentaux incomplets. Dans un second temps, les résultats d’essais réalisés lors de la thèse sont utilisés pour identifier les transferts thermiques au sein des turbocompresseurs. Un modèle dynamique d’échange thermique est ensuite construit. Le modèle final permet d’obtenir une cartographie complète des champs compresseur et turbine, tout en prenant en compte l’impact des transferts thermiques sur les performances des turbomachines. La seconde partie porte sur la modélisation des restrictions (soupapes, changements de section…) rencontrées sur les MCI. Une méthode pour la résolution quasistatique des équations Euler 1D complètes (masse, énergie, moment) est proposée afin de construire les cartographies associées aux restrictions considérées, sans introduire de coefficient de correction expérimental. Ces cartographies sont utilisables en tant que modèles moyens ou utilisées comme conditions limites de codes aérodynamiques 1D. En cas de couplage avec un code 1D, une méthode de correction d’entropie basée sur un algorithme de linéarisation est proposée afin de réduire le temps de convergence aux conditions limites du domaine de calcul.Finalement, les développements sont validés expérimentalement en écoulements stabilisés et transitoires. / The present thesis will focus on the modeling of the airpath components for the development of the control laws. The goal is to ensure the accuracy of the models while liming the associated calculation time. Indeed, models associated to the development of control laws must ensure law calculation time in order to be compliant with automotive development processes. In the same time, pollutant emissions regulations are increasingly restrictive, which has a direct impact on the required accuracy level of the models. In a first part, the physical equations of the radial turbocharger are introduced in a datamap-based approach. The fundamental equations of turbomachines are developed and processed in order to build an interpolation / extrapolation method for incomplete turbocharger datamaps. Then, measurements from tests realized during the thesis are analyzed in order to identify the heat exchanges occurring within turbochargers. A dynamical heat exchange model is then built. The final turbocharger model allows to build complete turbocharger datamaps, while taking into account the effect of thermal exchanges. The second part focuses on the modeling of ICE flow restrictions (valves, change of pipe area…). A method is proposed in order to solve the quasistatic formulation of the Euler 1D equation (mass, energy, momentum). This allows to build datamaps associated to the studied components, without having to introduce an experimental corrective coefficient. These datamaps can be used as mean-value models or as boundary conditions of a 1D numerical scheme. In case of coupling with a 1D scheme, an entropy correction scheme based on local linearization has been developed in order to improve the convergence speed at the boundary condition. Finally, proposed developments are validated experimentally under stabilized and transient flow conditions.

The interest in one dimensional (1D) magnetism has been strongly renewed with the synthesis of many magnetic compounds which exhibit a quasi one dimensional magnetic behaviour. One of the peculiarities of this 1D system is the absence of a long range magnetic ordered phase at any finite temperature for the ideal 1D system with short range interaction. Tetramethylammonium manganese chloride (CH3)4NMnCl3(TMMC) exhibits the properties of an ideal one dimensional antiferromagnets for temperature above 1 K, the transition to a three dimensional (3D) long range ordered state only occurs at 0.84K. In addition to its magnetic transition, TMMC exhibits structural phase transition due to the ordering of the tetramethylammonium (TMA) ions which makes also this compound very attractive from a lattice dynamical point of view. Structural phase transitions of tetramethylammonium manganese chloride (TMMC), tetramethylammonium manganese bromide (TMMB) and tetramethylammonium manganese chloride doped with 8% Cu (TMMC:Cu) of the hexagonal type compounds are investigated using the Brillouin scattering method. These crystals show pronounced acoustic anomalies in the region of the structural phase transition. The acoustic anomalies were observed by measuring sound velocity and hence the elastic constant can be deduced. The phase transition temperatures were observed at 129.6K and 388.6K (TMMC), 114.6K and 377.6K (TMMB) and at 108.6K and 359.6K (TMMC:Cu). The elastic constant at room temperature were C11 = 2.10 (TMMC) and C11 = 1.59 (TMMB) in units of 1010 Nm-2. The phase transition of these compounds were further investigated macroscopically using the Differential Scanning Calorimetry (DSC) method. Activation energies of TMMC, TMMC:Cu, TMMB and deuterated TMMB at the phase transition were determined using this method. The values are 70.612 kJ/mol (TMMC), 49.224 kJ/mol (TMMC:Cu), 51.747 kJ/mol (TMMB) and 69.909 kJ/mol (d12-TMMB). The elastic constant of the linear chain antiferromagnet CsNiCl3 and RbNiCl3 was also determined using the Brillouin scattering method. The room temperature measurements give C11 = 3.77 (3.71) and C33 = 5.62 (5.42) in units of 1010 Nm·2 for CsNiCl3 and RbNiCl3 respectively. The phonon dispersion curves at room temperature in the hexagonal CsFeBr3 have been studied using the inelastic neutron scattering technique. From the initial slope of the dispersion curve, the sound velocity was deduced which enable us to calculate the elastic constant of CsFeBr3 at room temperature. The values obtained are C11 = 7.33, C66 = 1.01, C33 = 2.58 and C44 = 0.56 in units of 1010 Nm·2.

Shock absorbers are crucial components of a vehicle’s chassis responsible for the trade-off between stability, handling, and passenger comfort. The aim of the thesis is to investigate the physical behavior of an advanced automotive racing shock absorber, known as TTR, developed by Öhlins Racing AB. This goal is achieved by developing a detailed lumped parameter numerical model of the entire TTR suspension in the advanced 1D simulation tool, AMESim. The shock absorber is mainly composed of the main cylinder with through-rod piston design and the gas reservoir located at the low pressure hydraulic line, which connects the compression and rebound sides. The mentioned sides are identical in terms of the components which are a High Speed Adjuster, a Low Speed Adjuster, and a check valve mounted in parallel. The adjusters are special hydraulic valves, which can be modified in terms of flow metering characteristics by means of external accessible screws. Adjustment is done in a series of discrete numbers called ‘clicks’. A fixed orifice and a spring-loaded poppet valve are responsible for controlling the piston low and high speed regions respectively. The developed AMESim numerical model is capable of capturing the physics behind the real shock absorber damping characteristics, under both static and dynamic conditions. The model is developed mainly using the standard AMESim mechanical, hydraulic and hydraulic component design libraries and allows discovering the impact of each single hydraulic component on the TTR overall behavior. In particular, the 1D model is presented in two levels of progressive physical complexity in order to improve the dynamic damping characteristics. Several physical phenomena are considered, such as the hydraulics volumes pressure dynamics, the contribution of external spring and pressure forces to the dynamic balance of the moving elements, the static and viscous frictions, and the elastic deformations induced by solid boundaries pressure. In this thesis, progressive model validation with different types of measurements is as well presented, covering the individual hydraulic components models as well as the entire shock absorber model. The measurements have been performed on the flow benches and dynamometers available at the Öhlins Racing measurements laboratory. These comparisons, deeply discussed in the thesis, allow discovering the impact of specific physical effects on the low and high speed hydraulic valves static performance and on the shock absorber dynamic behavior. Numerical results show good agreement, especially at low and medium frequencies and symmetric ‘click’ adjustments on compression and rebound sides. Further model development is necessary in the other areas, for example by considering more complex models of the valve dynamics and fluid flow patterns, i.e. flow forces, together with more advanced models of the sealing elements viscous friction, and thermal effects. Finally, the AMESim environments offered a good level of flexibility in designing the TTR hydro-mechanical system, by allowing the user to choose between different levels of model complexity.

Made available in DSpace on 2014-06-12T17:42:16Z (GMT). No. of bitstreams: 2 arquivo7111_1.pdf: 1388033 bytes, checksum: 645508c7ef101ae31fed17e866992644 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2011 / Faculdade de Amparo à Ciência e Tecnologia do Estado de Pernambuco / Na atualidade, metade da população mundial reside nos grandes centros urbanos. Grande parcela desse crescimento tem ocorrido em países em desenvolvimento. No Brasil, já se verifica um contingente de cerca de 80% da população residindo nas áreas urbanas. Neste crescente cenário de urbanização, impactos ambientais e socioeconômicos decorrentes de eventos hidrológicos têm sido recorrentes, afetando grande parte da população. O aumento da impermeabilização reduz as taxas de infiltração, que por sua vez leva à diminuição das taxas de recarga para os aquíferos e à diminuição do escoamento de base. O escoamento superficial é intensificado, aumentando em velocidade e, a frequência e magnitude dos picos de cheia, levando ocasionalmente às enchentes. Nesse contexto, têm sido empregados os telhados verdes em várias partes do mundo principalmente com finalidades estéticas de valorização do espaço urbano e para melhoria do conforto ambiental. Essas áreas verdes podem servir também para detenção do escoamento superficial, minimizando as enchentes urbanas. A simulação da dinâmica da água no solo do telhado verde realizada no programa Hydrus 1-D, a partir das características do sistema do teto verde implantado e dados obtidos em campo, proporcionou a caracterização da dinâmica da água em seu perfil de solo, fornecendo subsídios quanto ao desempenho deste dispositivo no amortecimento do escoamento superficial oriundo dos telhados

Modelos matemáticos são comumente utilizados no estudo da dinâmica da água no solo não-saturado. A principal equação para se quantificar esta dinâmica é a equação diferencial de Richards. Sua solução direta é impossível na maioria dos casos, necessitando de métodos numéricos, dos quais se destaca a utilização das funções de condutividade e de retenção de água de Van Genuchten - Mualem para obtenção da solução numérica. Diante disto, o objetivo deste trabalho foi apresentar um novo método para a obtenção dos parâmetros de Van Genuchten - Mualem, que utiliza a modelagem inversa de dados de teores de água medidos periodicamente e obtidos por experimento de evaporação assistido por atenuação de radiação gama, para a modelagem inversa foi utilizado o software Hydrus-1D. O método foi testado em amostras com diferentes texturas, colhidas em 11 localidades na região de Piracicaba-SP. Dos conjuntos de exemplares analisados, apenas um apresentou resultado insatisfatório, concluindo que o método é válido / Mathematical models are commonly used in studies of water dynamics in unsaturated soil. The main equation to quantify water the dynamics is the differential Richards equation. Its analytical solution is impossible in almost all cases, requiring numerical methods, among which the Van Genuchten - Mualem water conductivity and water retention functions are frequently used to obtain the numerical solution. The objective of this work is to present a new method to obtain the Van Genuchten - Mualem parameters, using the inverse modeling of water content data measured periodically by gamma radiation attenuation in evaporating samples. Hydrus-1D software was used for the inverse modeling. The method was tested in samples with different textures, collected in 11 locations in the region of Piracicaba, state of São Paulo, Brazil. Of the sets of samples analyzed, only one presented an unsatisfactory result, concluding that the method is valid

Les phénomènes quantiques ouvrent des possibilités nouvelles et révolutionnaires dans les domaines du calcul et de la cryptographie. Il est attendue, que des problèmes impossibles à résoudre avec des moyens classiques, peuvent être résolus par des ordinateurs quantiques, et la communication devient absolument sécurisée si elle est encodée dans un état de système quantique – un bit quantique. Un effort important a récemment été consacré à la recherche sur le transfert déterministe d’information entre photons et atomes, fonctionnant comme des bits quantiques volants et stationnaires respectivement. L’interaction entre ces deux composants est augmentée s’ils sont placés dans un milieu optique unidimensionnel, réalisant un système appelée “un atome 1D”. L’étude de ce milieu optique et des ses applications aux technologies quantiques constitue l’objectif de cette thèse.Tout d’abord, nous explorons l’interaction entre le champ lumineux et un atome 1D, en prenant une boîte quantique semi-conductrice dans un micropilier comme exemple. Nous étudions le contrôle cohérent de ce système avec des impulsions lumineuses afin de trouver un moyen optimal de contrôler son état, en faisant varier la puissance, la forme et la durée d’une impulsion, ainsi que la statistique de l’état quantique du champ lumineux. Nous étudions également l’impact de l’atome 1D sur l'état du champ réfléchi en fonction des paramètres du système expérimental.Nous poursuivons avec l’étude de l’état quantique du champ lumineux réfléchi, en nous concentrant sur sa pureté. C’est important pour transmettre fidèlement l’état superposition d’un bit stationnaire à un autre par la lumière, qui agit comme un bit quantique volant. Nous développons une méthode de caractérisation expérimentale de la pureté et l’appliquons à des données expérimentales, démontrant ainsi que la technologie moderne permet de créer des superpositions de haute pureté.Enfin, nous nous concentrons sur la lecture d’un qubit stationnaire basé sur un spin dans un environnement unidimensionnel. Nous étudions comment la lumière polarisée peut être utilisé pour cela, en montrant qu’il est possible de lire l’état de spin en détectant qu’un seul photon. Nous explorons différents écarts de ce régime optimal. Nous étudions également la décohérence de l’état de spin due à l’interaction avec le champ lumineux, et back-action de la mesure, montrant que l’état de spin peut être “gelé”. C’est une manifestation de l’effet Zeno quantique, qui permet la préparation du qubit dans un état arbitraire. Cela ouvre des perspectives pour la réalisation efficace de bits quantiques stationnaires basés sur des spins uniques incorporés dans un environnement électromagnétique unidimensionnel. / Quantum phenomena give rise to new and revolutionary possibilities in the fields of computation and cryptography. The problems that are unsolvable with classical means are expected to be solved by quantum computers, and communication becomes absolutely secure, if it is encoded in a state of a quantum system. A large effort has been recently paid to research of deterministic transfer of information between photons and atoms, acting as flying and stationary quantum bits respectively. The interaction between these two components is enhanced, if they are put in a unidimensional medium, realizing a so called "1D atom". The study of this specific optical medium and its applications to quantum technologies constitutes the objective of this thesis.First, we explore the light-matter interface realized as a 1D atom, with a semiconductor quantum dot in a micropillar cavity as an example. We study the coherent control of this system with light pulses in order to find an optimal way to control its state, varying the power, shape and duration of a pulse and statistics of the state of light field. We also study the impact of the 1D atom on the state of the reflected field as a function of parameters of the experimental device, describing the filtering of single photon Fock state from incident pulse.We continue with the study of the quantum state of the scattered light field, focusing on its purity. This is required to faithfully transmit the superposition state of one stationary qubit to another using light as a flying quantum bit. We develop a method to experimentally characterize the purity, and apply it to experimental data, showing that the state of art technology allows to create high-purity superpositions.Finally, we focus on the readout of a stationary qubit based on a single spin in a unidimensional environment. We study how to efficiently use polarized light for this purpose, showing that it is possible to readout the spin state, by detection of only one photon. We explore different deviations from this optimal regime. We also study the decoherence of the spin state due to interaction with the light field and the back-action of the measurement, showing that it is possible to freeze the spin state due to the quantum Zeno effect, which allows the preparation of the qubit, based on it, in an arbitrary superposition state. This opens perspectives towards efficient realization of stationary quantum bits based on single spins embedded in unidimensional electromagnetic environment.

Cette thèse porte sur la modélisation 0D/1D de la combustion Diesel dans les moteurs récents. L’objectif est d’augmenter la précision des modèles tout en limitant les temps de calcul associés afin d’utiliser la simulation comme un outil dédié à la mise au point. Dans une première partie, le développement d’un modèle 0D orienté simulation système est présenté. La prise en compte de l’ensemble des phénomènes physico-chimiques se déroulant dans la chambre de combustion confère au modèle un niveau de prédictivité conséquent. Un nouveau modèle de combustion de prémélange est proposé, permettant une modélisation détaillée des combustions fortement diluées et des combustions relatives aux injections précoces. Une approche innovante permettant de quantifier les interactions entre les jets pour la multi injection est également proposée. Après calibration sur un nombre restreint d’essais moteur, les résultats du modèle global sont comparés à des mesures expérimentales pour toute la plage de fonctionnement du moteur. La seconde partie de ce travail porte sur la modélisation 1D de la combustion Diesel. Un modèle de jet Diesel est d’abord développé et validé sur des mesures expérimentales. Ce modèle est ensuite étendu à des conditions réactionnelles à l’aide d’un couplage avec un modèle de combustion. Ce dernier s’appuie sur une tabulation des mécanismes de cinétique chimique, ainsi que sur une approche Eddy Break-Up permettant de modéliser le taux de réaction lié au micro mélange. Ce modèle est ensuite intégré à un modèle de chambre de combustion et une première validation du modèle sur des essais moteur réels est entreprise. / The present thesis focuses on the 0D/1D Diesel combustion modeling of recent engines. The goal is to improve models accuracy while minimizing computation times in order to use simulation as a tool for engine pre-mapping. In the first part, a 0D model designed as a system simulation-oriented tool is proposed. The main contribution of this study is the modeling of the premixed part of the Diesel combustion. This model allows a detailed modeling of highly diluted combustion and combustion related to early injections. A new approach to quantify interactions between each spray in the case of multi injection strategies is also proposed. After calibration using a very small number of engine tests, results for the global combustion chamber model are compared with experimental measurements for the overall engine operating conditions. The second part of this work deals with the 1D Diesel combustion modeling. A Diesel spray model is at first developed and validated on experimental measurements. This model is then extended to reaction conditions using the coupling with a combustion model. The combustion model makes use of tabulated local reaction rates of fuel and is based on the Eddy Break-Up approach to describe the reaction rate related to the turbulent mixing process. The next step is the integration of the burning spray model into a Diesel engine combustion chamber model. A first validation using experimental results for a recent Diesel engine is done.

Modelos matemáticos são comumente utilizados no estudo da dinâmica da água no solo não-saturado. A principal equação para se quantificar esta dinâmica é a equação diferencial de Richards. Sua solução direta é impossível na maioria dos casos, necessitando de métodos numéricos, dos quais se destaca a utilização das funções de condutividade e de retenção de água de Van Genuchten - Mualem para obtenção da solução numérica. Diante disto, o objetivo deste trabalho foi apresentar um novo método para a obtenção dos parâmetros de Van Genuchten - Mualem, que utiliza a modelagem inversa de dados de teores de água medidos periodicamente e obtidos por experimento de evaporação assistido por atenuação de radiação gama, para a modelagem inversa foi utilizado o software Hydrus-1D. O método foi testado em amostras com diferentes texturas, colhidas em 11 localidades na região de Piracicaba-SP. Dos conjuntos de exemplares analisados, apenas um apresentou resultado insatisfatório, concluindo que o método é válido / Mathematical models are commonly used in studies of water dynamics in unsaturated soil. The main equation to quantify water the dynamics is the differential Richards equation. Its analytical solution is impossible in almost all cases, requiring numerical methods, among which the Van Genuchten - Mualem water conductivity and water retention functions are frequently used to obtain the numerical solution. The objective of this work is to present a new method to obtain the Van Genuchten - Mualem parameters, using the inverse modeling of water content data measured periodically by gamma radiation attenuation in evaporating samples. Hydrus-1D software was used for the inverse modeling. The method was tested in samples with different textures, collected in 11 locations in the region of Piracicaba, state of São Paulo, Brazil. Of the sets of samples analyzed, only one presented an unsatisfactory result, concluding that the method is valid

The master’s thesis deals with the Wankel rotary engines and their 1D simulations using a thermodynamic simulation software for the piston engines. The necessary steps for creation of the equivalent model of the four-stroke three-cylinder combustion engine are provided. The engine used for the validation model was Aixro XR 50. The data measured on this engine during testing were used to validate the created thermodynamic model. The discharge coefficient calculation of the intake and the exhaust ports is shown. The 11kW engine design is created using validated thermodynamic model.

The master’s thesis deals with the Wankel rotary engines and their 1D simulations using a thermodynamic simulation software for the piston engines. The necessary steps for creation of the equivalent model of the four-stroke three-cylinder combustion engine are provided. The engine used for the validation model was Aixro XR 50. The data measured on this engine during testing were used to validate the created thermodynamic model. The discharge coefficient calculation of the intake and the exhaust ports is shown. The 11kW engine design is created using validated thermodynamic model.