Les travaux présentés dans cette thèse concernent l'étude thermique des batteries Li-ion en vue d'une application de gestion thermique pour l'automobile. La compréhension des phénomènes thermiques à l'échelle accumulateur est indispensable avant de réaliser une approche de type module ou pack batterie. Ces phénomènes thermiques sont mis en évidence à partir d'une modélisation thermique globale de deux accumulateurs de différentes chimies, en décharge à courant constant. La complexité du caractère résistif de l'accumulateur Li-ion a mené au développement d'un modèle prenant en compte l'interaction entre les phénomènes électrochimiques et thermiques, permettant une approche prédictive de son comportement. Enfin la réalisation de deux boucles expérimentales, de simulation de systèmes de gestion thermique d'un module de batterie, montre les limites d'un refroidissement classique par air à respecter les critères de management thermique. En comparaison, le second système basé sur l'intégration innovante d'un matériau à changement de phase (MCP) se montre performant lors de situations usuelles, de défauts ou encore lors du besoin d'une charge rapide de la batterie. / This work relates to the thermal study of Li-ion batteries in order to develop an optimized battery thermal management system. The understanding of thermal phenomena at cell scale is essential before to undertake an approach of the battery module or pack. Galvanostatic discharges of two kind of Li ion cells are modeled to highlight thermal phenomena. The complexity of the resistive behavior of Li-ion cell led to the development of an electrochemical-thermal coupled model to get a predictive approach. Then, two experimental tests benches were designed so as to compare two battery thermal management systems (BTMS). Restrictions of air cooling highlight its disability to achieve thermal management criteria. Innovative integration of a phase change material (PCM) was then tested under several uses of the battery module. This new BTMS showed really promising performances during intensive driving cycles, failure tests, and when a fast charge is needed.

La recherche dans le domaine de l’informatique neuro-inspirée suscite beaucoup d'intérêt depuis quelques années. Avec des applications potentielles dans des domaines tels que le traitement de données à grande échelle, la robotique ou encore les systèmes autonomes intelligents pour ne citer qu'eux, des paradigmes de calcul bio-inspirés sont étudies pour la prochaine génération solutions informatiques (post-Moore, non-Von Neumann) ultra-basse consommation. Dans ce travail, nous discutons les rôles que les différentes technologies de mémoire résistive non-volatiles émergentes (RRAM), notamment (i) Phase Change Memory (PCM), (ii) Conductive-Bridge Memory (CBRAM) et de la mémoire basée sur une structure Metal-Oxide (OXRAM) peuvent jouer dans des dispositifs neuromorphiques dédies. Nous nous concentrons sur l'émulation des effets de plasticité synaptique comme la potentialisation à long terme (Long Term Potentiation, LTP), la dépression à long terme (Long Term Depression, LTD) et la théorie STDP (Spike-Timing Dependent Plasticity) avec des synapses RRAM. Nous avons développé à la fois de nouvelles architectures de faiblement énergivore, des méthodologies de programmation ainsi que des règles d’apprentissages simplifiées inspirées de la théorie STDP spécifiquement optimisées pour certaines technologies RRAM. Nous montrons l’implémentation de systèmes neuromorphiques a grande échelle et efficace énergétiquement selon deux approches différentes: (i) des synapses multi-niveaux déterministes et (ii) des synapses stochastiques binaires. Des prototypes d'applications telles que l’extraction de schéma visuel et auditif complexe sont également montres en utilisant des réseaux de neurones impulsionnels (Feed-forward Spiking Neural Network, SNN). Nous introduisons également une nouvelle méthodologie pour concevoir des neurones stochastiques très compacts qui exploitent les caractéristiques physiques intrinsèques des appareils CBRAM. / Research in the field of neuromorphic- and cognitive- computing has generated a lot of interest in recent years. With potential application in fields such as large-scale data driven computing, robotics, intelligent autonomous systems to name a few, bio-inspired computing paradigms are being investigated as the next generation (post-Moore, non-Von Neumann) ultra-low power computing solutions. In this work we discuss the role that different emerging non-volatile resistive memory technologies (RRAM), specifically (i) Phase Change Memory (PCM), (ii) Conductive-Bridge Memory (CBRAM) and Metal-Oxide based Memory (OXRAM) can play in dedicated neuromorphic hardware. We focus on the emulation of synaptic plasticity effects such as long-term potentiation (LTP), long term depression (LTD) and spike-timing dependent plasticity (STDP) with RRAM synapses. We developed novel low-power architectures, programming methodologies, and simplified STDP-like learning rules, optimized specifically for some RRAM technologies. We show the implementation of large-scale energy efficient neuromorphic systems with two different approaches (i) deterministic multi-level synapses and (ii) stochastic-binary synapses. Prototype applications such as complex visual- and auditory- pattern extraction are also shown using feed-forward spiking neural networks (SNN). We also introduce a novel methodology to design low-area efficient stochastic neurons that exploit intrinsic physical effects of CBRAM devices.

L’objectif général de la thèse est de contribuer à l’amélioration de nos connaissances sur les métamorphoses de la neige sèche et sur sa description physique, à l’échelle microscopique (grains de glace et pores) et macroscopique (couche de neige). Dans un premier temps,la méthode d’homogénéisation basée sur les développements asymptotiques à échelles multiples est appliquée à la physique des métamorphoses de la neige sèche. On présente ainsi les descriptions macroscopiques équivalentes du transport de vapeur et de chaleur dérivées à partir de la description de la physique à micro-échelle. On considère à l’échelle des grains la diffusion, la conduction, et la convection forcée, couplées aux changements de phase (sublimation et déposition). Dans un second temps, les propriétés effectives de transport impliquées dans les descriptions macroscopiques (conductivité thermique effective, coefficient effectif de diffusion de vapeur et perméabilité intrinsèque) sont estimées à l’aide d’images 3D de neige couvrant toute la gamme de masse volumique et de types de neige. Enfin, on s’intéresse au suivi temporel des métamorphoses. Les liens entre la microstructure et les propriétés effectives d’une couche de neige sont mis en évidence au cours d’une métamorphose de gradient de température en utilisant des images 3D.On présente ensuite une cellule cryogénique que nous avons mise au point pour le suivi grains à grains par tomographie des évolutions d’un échantillon de neige au cours des métamorphoses, et qui s’utilise à température ambiante. / The main objective of the thesis is to improve our knowledge about dry snow metamorphismand its physical description, at the microscopic (ice grains and pores) andmacroscopic (snow layer) scales. First, the homogenization method of multiple scaleexpansions is applied for the first time to the physics involved in dry snow metamorphism.This way, we present the equivalent macroscopic descriptions of heat and vaportransfers derived from the physical description at micro-scale. We consider at the grainscale diffusion, conduction, and forced convection, coupled to phase changes (sublimationand deposition). Second, the effective properties of transport arising in the macroscopicdescriptions (effective thermal conductivity, effective coefficient of vapor diffusion, andintrinsic permeability) are estimated from 3D images of snow spanning the whole range ofdensity and snow types. Finally, the monitoring of metamorphism with time is considered.The relationship between the microstructure and the effective properties of a snow layerare investigated during temperature gradient metamorphism using 3D images. We presentthen a new cryogenic cell that we developed to monitor the grain to grain evolution of asnow sample by time-lapse tomography during the metamorphism, and which operates atroom temperature.

O elemento hélio, encontrado principalmente em reservas de gás natural, entra em condensação à temperatura de 4,2K, e é a única substância conhecida que permanece no estado líquido até o zero absoluto. Na fase liquida, o hélio apresenta ainda, em K, outra mudança de fase, onde passa de líquido comum à superfluido, com viscosidade praticamente nula. Estas propriedades conferem ao hélio importantes aplicações. Uma hdas principais aplicações é como agente refrigerante em supercondutores, como por exemplo, no acelerador de partículas LHC, que está sendo construído na fronteira da França com a Suíça, em aparelhos de ressonância magnética, satélites artificiais, etc. Neste trabalho, são apresentados dois modelos matemáticos para a transferência de calor no hélio líquido. O primeiro modelo, considerando apenas movimentos macroscópicos, é derivado com base nas leis constitutivas de Fourier e de Gorter-Mellink. O segundo modelo, baseado nas técnicas de Fremond, inclui movimentos microscópicos e pode ser visto como uma regularização do primeiro modelo. Os dois modelos são governados por equações diferenciais fortemente não lineares resultantes da não linearidade da lei de Gorter-Mellink e da mudança de fase. Ambos os modelos podem ser considerados casos particulares do problema de Stefan de duas fases, sendo que em uma das fases o fluxo de calor é governado pela equação não-linear do problema conhecido como p-laplaciano, com p=4/3. São também apresentadas técnicas para resolver de forma eficiente o problema do p-laplaciano, tanto para valores grandes de p, p>>2, quanto para valores de p próximos à 1, que constituem importantes desafios numéricos. Para tanto são propostos dois métodos iterativos simples, um baseado no método de quase-Newton, com termo de relaxação e, outro através da decomposição de Helmholtz, gerando um sistema de equações cujas matrizes são constantes, o que diminui significativamente o custo computacional. Experimentos numéricos são realizados para testar a eficiência dos modelos numéricos propostos bem como dos algoritmos desenvolvidos para resolver os sistemas de equações algébricas não lineares resultantes das aproximações por elementos finitos. São apresentados resultados de estudos de convergência, mostrando taxas de convergência ótimas ou quase ótimas, comparáveis às das interpolantes. Para o problema com mudança de fase, devido à descontinuidade do gradiente da temperatura sobre a interface que separa as duas fases do hélio líquido, as taxas de convergência não são ótimas. Usando malhas adaptativas, consegue-se taxas ótimas também para o problema com mudança de fase. Usando dados experimentais, encontrados na literatura, para os parâmetros de condutividade térmica, densidade e calor específico, dependentes da temperatura, são também apresentados testes de validação do modelo e exemplos de possíveis aplicações. Nos testes de validação do modelo, compara-se a solução numérica do modelo matemático com resultados experimentais para a temperatura, encontrados na literatura. / The element helium, found mainly in natural gas reserves, condenses at temperature of 4.2K, and is the unique known substance that remains in liquid to absolute zero. In the liquid phase, the helium presents still another phase change in 2.19K, where passes of common liquid to superfluous liquid, with almost zero viscosity. These properties give the helium important applications. One of the major applications is as a coolant in superconductors, such as in the particle accelerator LHC, which is being built in the French border with Switzerland, in magnetic resonance devices, artificial satellites, etc.. In this paper, we present two mathematical models for heat transfer in liquid helium. The first model, considering only macroscopic movements, is derived based on constitutive laws of Fourier and Gorter-Mellink. The second model, based on techniques of Fremond, includes microscopic movements and can be seen as a regularization of the first model. Both models are governed by highly nonlinear differential equations resulting from the nonlinearity of the law of Gorter-Mellink and change of phase. Both models can be considered special cases of the Stefan problem in two phases, with phase one of the heat flux is governed by non-linear equation of the problem known as p-Laplacian, with p = 4/3. We also presented techniques to efficiently solve the problem of p-Laplacian, both for large values of p, p>> 2, and for values of p close to 1, which are major numerical challenges. Are proposed two simple iterative methods, one based on the method of quasi-Newton, with the relaxation term and the other by the Helmholtz decomposition, creating a system of equations whose matrices are constant, which reduces significantly the computational cost. Numerical experiments are conducted to test the efficiency of numerical models proposed and the algorithms developed for solving systems of nonlinear algebraic equations arising from approximations by finite elements. Are also presented results of studies of convergence, showing rates of optimal or near optimal convergence, comparable to that of interpolates. For the problem with phase change, due to the discontinuity of the gradient of temperature on the interface separating the two phases of liquid helium, the rate of convergence is not optimal. Using adaptive mesh, it is also great rates to the problem with change of phase. Using experimental data found in literature, for the parameters of thermal conductivity, density and specific heat, temperature dependent, are also presented for validation testing of the model and examples of possible applications. In tests for validating the model, compared to the numerical solution of the mathematical model with experimental results for the temperature found in literature.

Icing has since long been identified as a serious issue in the aeronautical world. Ice build up, due to the presence of supercooled water droplets in clouds, leads to degradation of aerodynamic and/or air intake performances, among other undesirable consequences. Hence aircraft manufacturers must comply with certifications and regulations regarding flight safety in icing conditions. In order to do so, ice protection systems are used. Due to the multi-physical context within which these systems operate, numerical simulation can be a valuable asset. The present work deals with the numerical modelling of electro-thermal ice protection systems. It is built around the development of three modules. Two of them are dedicated to modelling heat transfer in the system and in the ice block. The other one models the mechanical behaviour of ice and fracture. Hence, the mechanical properties of atmospheric ice are reviewed in order to identify some mechanical parameters relevant to the fracture model. The fracture mechanics numerical method is used to investigate possible ice shedding mechanisms, that is to say the mechanisms leading to the detachement of ice, which are not yet well understood. The final goal of this work is to propose a completely coupled 2nd generation simulation methodology for electro-thermal ice protection systems. Hence the feasibility of a coupled thermal computation with ice shedding prediction based on the developed modules is shown. / Le givrage a depuis longtemps été identifié comme une problématique sérieuse dans le monde aéronautique.L’accrétion de givre, due à la présence de gouttelettes d’eau surfondue dans les nuages, dégrade les performances aérodynamiques et le rendement des entrées d’air parmi d’autres conséquences néfastes. Ainsi, les avionneurs sont sujets à des règles de certifications concernant la capacité à voler en conditions givrantes. Pour se faire, des systèmes de protection contre le givre sont utilisés. En raison de la complexité des phénomènes physiques mis en jeux, la simulation numérique constitue un atout lors de la phase de conception. Ce travail de thèse porte sur la modélisation et la simulation numérique des systèmes électrothermiques de protection contre le givre. Il s’articule autour de trois approches de modélisation, qui ont donné lieu au développement de trois modules. Deux d’entre eux sont dédiés à la simulation du transfert de chaleur dans le système et dans la glace (changement de phase). Le troisième est lié à la modélisation du comportement mécanique du givre atmosphérique avec fissuration. Ainsi, les propriétés mécaniques du givre atmosphérique sont revues de façon à pouvoir identifier les paramètres intervenant dans le modèle de fissuration. Ce modèle est ensuite utilisé pour étudier les mécanismes possibles de détachement du givre, qui ne sont à l’heure actuelle pas encore bien compris. Le but final de ce travail est de proposer une méthodologie de simulation couplée pour les systèmes électrothermiques de protection contre le givre. Ainsi, la faisabilité d’un calcul couplé thermique-fissuration avec prédiction de détachement de givre est présentée.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-11-07T12:20:00Z No. of bitstreams: 1 gessicalacerdasiqueira.pdf: 2914379 bytes, checksum: 13b701ceac7f6562db4c0a3167e60c32 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-11-09T14:37:17Z (GMT) No. of bitstreams: 1 gessicalacerdasiqueira.pdf: 2914379 bytes, checksum: 13b701ceac7f6562db4c0a3167e60c32 (MD5) / Made available in DSpace on 2017-11-09T14:37:17Z (GMT). No. of bitstreams: 1 gessicalacerdasiqueira.pdf: 2914379 bytes, checksum: 13b701ceac7f6562db4c0a3167e60c32 (MD5) Previous issue date: 2017-07-13 / Processos de transferência de calor com mudança de fase são de interesse em várias áreas da ciência, engenharia e aplicações industriais. Em processos de solidificação, por exemplo, a ocorrência de trincas e formações de vazios causados por uma redução na capacidade de resistir a esforços mecânicos geralmente são observados, e é de grande importância prever esses comportamentos para que possam ser evitados. Uma das formas de se estudar esses fenômenos e suas aplicações é através do uso de modelos matemáticos que possibilitam, através do uso de métodos de discretização temporal e espacial a realização de simulações computacionais. Dessa forma, este trabalho estudou e comparou diferentes formulações numéricas do método dos elementos finitos para problemas de transferência de calor com mudança de fase levando em consideração a condução como mecanismo de transferência de calor. Na metodologia usada, os processos de solidificação ou fusão foram tratados por meio de métodos baseados em malhas fixas. Diversos experimentos numéricos foram realizados para se analisar a adequabilidade dos métodos estudados e propostos neste trabalho e os resultados obtidos foram satisfatórios. / Heat transfer processes with phase change are of great interest in several areas of science, engineering and industrial applications. In some of these applications, such as in solidification processes, the occurrence of cracks and void formation caused by a reduction in the ability of the material to resist mechanical loadings are generally observed, and it is of great importance to predict these behaviors so that they can be circumvented. One of the ways to study these phenomena and their applications is through the use of mathematical models and executing of computer simulations. Thus, this work studied and compared different numerical formulations of the finite element method for the heat transfer problems with phase change considering only conduction as the main mechanism of heat transfer. In the methodology used, the solidification or melting processes were treated by means of numerical methods based on fixed meshs. Several numerical experiments were performed to analyze the suitability of the methods studied and proposed in this work and the results were satisfactory.

[pt] A pasta de gelo é uma mistura de água e um aditivo, com finas partículas de gelo, apresentando uma alta densidade de energia térmica. O principal motivo de sua utilização deve-se à combinação do aproveitamento do calor latente na mudança de fase com capacidade de ser bombeado. O presente trabalho trata do estudo experimental sobre a transfe rência de calor e queda de pressão, com mudança de fase, utilizando uma mistura de propileno glicol - água com 13,8% de concentração em peso, num trocador de calor de placas com arranjo em U para 16 placas. Realizaram-se testes de troca de calor com escoamento em paralelo e contra-corrente, para duas condições de fração mássica de gelo e números de Reynolds para a pasta de gelo entre 150 e 425, com diferentes condições de carga térmica. Dos testes foram observados aumentos de até 25% no coeficiente global de troca calor, ao se incrementar a vazão e, conseqüentemente, o número de Reynolds da pasta de gelo. Com o aumento da fração de gelo melhora-se a capacidade de resfriamento, diminuindo o número de Nusselt da pasta de gelo. O coeficiente global de troca, porém, começa a diminuir. Na literatura esta relação ainda não está bem definida. Alguns autores relatam ganhos, perdas ou indiferença no coeficiente global. Resultados do coeficiente global e do número de Nusselt, nos modos paralelo e contra-corrente, não apresentaram grande diferença. A capacidade de resfriamento em contra-corrente foi maior, apresentando valores de efetividade cerca 10% acima dos resultados observados no modo paralelo. Os fatores de atrito encontrados variaram entre 0,030 a 0,085, o que concorda com os resultados apresentados por outros pesquisadores. Como era de se esperar, o fator de atrito diminui com o aumento da vazão mássica e de maneira inversa com o aumento da fração de gelo. / [en] Ice slurry is an aqueous solution from which ice crystals are formed. These crystals possess high energy density, in the form of latent heat. Furthermore, the fact of being a slurry makes it an excellent energy carrier, for it can be easily pumped. The present work presents an experimental procedure to assess the heat transfer and the pressure loss, with phase change, using a mixture of propyleneglycol and water, 13.8% weight, in a U-plate heat exchanger with 16 plates. Tests were carried out for both parallel and counter-flow configurations of the heat exchanger, for two different mass fractions of ice and with Reynolds numbers between 150 and 425 for the slurry. Different thermal loads were considered. The experiments showed increases of up to 25% in the overall heat transfer coeficiente when the slurry flow is increased, with consequent increase in the Reynolds number. By increasing the ice fraction the cooling capacity is improved, reducing the Nusselt number. In the literature, this relationship is not clearly defined, as some authors show gains in the heat exchange coefficient, losses, and some others did not observe any influence on this parameter. Regarding the heat Exchange configuration, i.e., parallel or counter-flow, neither the heat exchange coefficient, nor the Nusselt number suffered major influence. The effectiveness is up to 10% higher for the counter-flow configuration in comparison with the parallel configuration. The figures found for the friction coefficient, between 0,030 and 0,085, are in good agreement with the literature. As expected, the friction coefficient decreases when the mass flow is increased, and increases when the ice fraction is increased.

L’amortisseur thermique étudié dans le cadre de cette thèse a pour objectif de limiter les pics de température des composants électroniques fonctionnant en régime transitoire au moyen d’une structure composite consistant en un réseau de nanotubes de carbone (NTC) rempli de matériau à changement de phase (MCP) solide-liquide, le tout étant contenu dans un boîtier en silicium (Si). Ce système passif vise à augmenter l’inertie thermique volumique du composant grâce à la chaleur latente du MCP tout en maintenant une bonne conductance thermique grâce aux NTC. Un dispositif expérimental polyvalent a été développé spécifiquement pour caractériser les différentes générations d’échantillons fabriqués par les partenaires du projet THERMA3D. L’excitation thermique de l’échantillon est réalisée au moyen d’un laser en face amont et la réponse thermique est mesurée par caméra infrarouge simultanément sur les faces amont et aval. L’application d’une peinture sélectionnée sur l’échantillon permet d’accéder à sa température après un étalonnage dédié. Des méthodes d’estimation de paramètres ont été développées pour quantifier les deux caractéristiques essentielles de l’amortisseur thermique que sont sa capacité de stockage thermique et sa résistance thermique. Les sensibilités de la résistance thermique aux caractéristiques de la connexion Si/NTC et à la longueur des NTC ont été étudiées et les résistances thermiques d’interface Si/NTC ont été identifiées comme dominantes au sein du système. Des essais de cyclage thermique ont permis d’évaluer la fiabilité de l’ensemble de manière accélérée. Le comportement du MCP et la qualité du matériau de scellement ont été analysés par voie optique. Par ailleurs, la plus élevée des deux résistances thermiques d’interface Si/NTC a été localisée grâce à la visualisation infrarouge du réseau de NTC à travers le silicium semi-transparent. Enfin, une méthode de contrôle non destructif de la qualité de l’interface Si/NTC a été développée pour les amortisseurs thermiques de dernière génération. / The purpose of the studied thermal damper is to smooth the temperature peaks of transient electronic components via a composite structure consisting of an array of carbon nanotubes (CNT) filled with solid-liquid phase change material (PCM), the whole being embedded in a silicon (Si) casing. This passive system is intended to increase the thermal inertia per unit of volume of the electronic component thanks to the latent heat of the PCM while maintaining a high thermal conductance thanks to the CNT. A versatile test bench was specifically developed in order to characterize the different generations of samples fabricated by the partners of the THERMA3D project. The thermal excitation of the front side of the sample is generated by a laser and the thermal response is measured simultaneously on the front and back sides by an infrared camera. A selected paint can be deposited on the sample in order to access its temperature by means of a dedicated calibration. Parameter estimation methods were developed in order to quantify both main characteristics of the thermal damper: its heat storage capacity and its thermal resistance. The sensitivities of the thermal resistance to the features of the Si/CNT connection and to the length of the CNT were studied and it was found out that the interfacial thermal resistances Si/CNT are dominant in the system. Thermal cycling tests enabled to assess the reliability of the thermal damper in an accelerated manner. The behavior of the PCM and the quality of the sealing material were optically analyzed. Besides, the infrared visualization of the CNT array through the semi-transparent silicon enabled to identify the highest of both Si/CNT interfacial thermal resistances. Finally, a non-destructive testing method for the evaluation of the quality of Si/CNT interfaces was developed for the latest generation of thermal dampers.

A ce jour, les solutions de stockage d'énergie apparaissent comme des solutions pertinentes permettant d'atteindre les cibles énergétiques futures et de répondre aux exigences environnementales actuelles. Le but de cette thèse est d’optimiser un système de stockage d'énergie thermique innovant basé sur un échangeur eau – matériaux à changement de phase. Ce système est couplé à l’architecture énergétique multi-sources d’un îlot composé de trois bâtiments à énergie positive situé à Lyon : l’îlot HIKARI. Afin de disposer d’un outil numérique robuste pour pouvoir optimiser cette technologie, un modèle numérique du système de stockage d’énergie thermique a été développé dans le but de reproduire le comportement du système de stockage de référence. Une fois fini, le modèle a été validé en trois étapes: une numérique et deux expérimentales. Dans un premier temps il a été validé numériquement, en comparant ses résultats avec un modèle conçu en adoptant une approche numérique différente (« Computational Fluid Dynamics »), dans un second temps il a été validé à l’échelle réelle en exploitant les données in-situ du système de HIKARI. Enfin, le modèle numérique a été validé expérimentalement grâce à un prototype expérimental conçu et réalisé à l’ENTPE dans le cadre des travaux de cette thèse reproduisant le comportement du système de stockage étudié. Après avoir été validé, le modèle a été utilisé pour procéder à l’optimisation de sa performance en utilisant la technique des algorithmes génétiques. L’analyse des résultats de ces simulations a notamment abouti sur des recommandations de dimensionnement et d’usage pour l’Ilot HIKARI et des bâtiments futurs intégrant la technologie de stockage étudiée. La thèse a été financée par l’Agence de l’environnement et de la maîtrise de l’énergie (ADEME) dans le cadre du projet « Optimisation des architectures Énergétiques multi-sources couplées aux techniques avancées du stockage d'énergie dans le bâtiment » en partenariat avec Bouygues immobilier et Manaslu – CMDL. / One of the most promising technics used in building applications for energy efficiency purposes is the thermal energy storage (TES). Despite the thorough research on TES techniques of the last years, the release to market of cost effective technologies is quite recent. The aim of this study is to optimize the energetic behavior of an innovative TES technology consisting on a water/PCM exchanger that is part of the multi-energy production and storage systems of HIKARI, a positive energy district located in Lyon and consisting of three buildings. In order to optimize this innovative technic, a numerical model reproducing the functioning of the reference system was created. In order to make a numerical validation a second numerical model was developed using a different software based on a different numerical method and, once the in situ data obtained from the reference system monitoring, a first experimental validation was obtained. Subsequently, an innovative experimental prototype reproducing the behavior of the reference PCM-Water heat exchanger has been realized, in order to validate and calibrate the numerical model and carry out a large amount of operating scenarios. Once the model numerically and experimentally validated, the optimization of the HIKARI’s cold storage system technology has been obtained using Genetic Algorithms (GAs) finding the best values to allocate to four characteristics of the cold storage system, in order to minimize two predefined objective functions linked to its functioning. This work was supported by the French Agency for Environment and Energy Management (ADEME) and it was part of the project “Optimization of innovative energy storage technologies when coupled to multi-sources energy architectures”, in cooperation with Bouygues immobilier and Manaslu – CMDL.

Ces travaux de recherche portent sur la conception et la réalisation de commutateurs RF basées sur l’intégration de matériaux innovants fonctionnels tels que le dioxyde de vanadium (VO2) et les alliages de chalcogénures de types Ge2Sb2Te5 (GST) et GeTe. Le principe de fonctionnement de ces composants repose sur le changement de résistivité que présentent ces matériaux. Le VO2 possède une transition Isolant-Métal (MIT) autour de 68°C à travers laquelle le matériau passe d’un état isolant (forte résistivité) à un état métallique (faible résistivité). La transition MIT présente l’intérêt de pouvoir être initiée sous l’effet de plusieurs types de stimuli externes (thermique, électrique et optique) avec de faibles temps de commutation. Les alliages de types GST et GeTe ont la particularité de commuter réversiblement entre un état amorphe à forte résistivité à un état cristallin à faible résistivité suite à un traitement thermique spécifique. Les commutateurs à base de GST ou de GeTe présentent l'avantage de pouvoir opérer en mode bistable car le changement de résistivité présenté par ces matériaux est de type non volatile. Les composants réalisés ont de bonnes performances électriques (isolation et pertes d’insertion) sur une large bande. Nos travaux de recherche visent à proposer une solution alternative aux solutions classiques (semi-conducteurs et MEMS-RF) pour réaliser des commutateurs RF qui peuvent être par la suite utilisés dans la conception des dispositifs reconfigurables (filtres, Antennes). / This research work focuses on the design and realization of RF switches based on the integration of new materials such as vanadium dioxide (VO2), Ge2Sb2Te5 (GST) and GeTe chalcogenides alloys. The operating principle of these devices is based on the resistivity change presented by these materials. VO2 exhibits a Metal-Insulator transition (MIT) around 68°C for which the material changes from an insulating state (high resistivity) to a metallic one (low resistivity). The MIT transition can be triggered in different ways (thermally, electrically and optically) with low switching time. GST and GeTe alloys have the particularity to be reversibly switched between a high resistive-amorphous state to low resistive-crystalline state, under a specific heat treatment. Thanks to the non-volatile resistivity change presented by these materials, GST/GeTe-based switches are able to operate in bistable mode. The fabricated devices exhibit good electrical performances (insertion loss and isolation) over a broadband. The aim of our work is to propose an alternative solution to conventional technologies (semiconductors and RF-MEMS), to design RF switches that can be used afterward in the design of reconfigurable devices (filters, antennas).