In wireless communication systems, multiple standards have been implemented to meet the past and present demands of different applications. This proliferation of wireless standards, operating over multiple frequency bands, has increased the demand for radio frequency (RF) components, and consequently power amplifiers (PA) to operate over multiple frequency bands. In this research work, a systematic approach for the synthesis of a novel dual-band matching network is proposed and applied for effective design of PA capable of maintaining high power efficiency at two arbitrary widely spaced frequencies. The proposed dual-band matching network incorporates two different stages. The first one aims at transforming the targeted two complex impedances, at the two operating frequencies, to a real one. The second stage is a dual-band filter that ensures the matching of the former real impedance to the termination impedance to 50 Ohm. Furthermore, an additional transmission line is incorporated between the two previously mentioned stages to adjust the impedances at the second and third harmonics without altering the impedances seen at the fundamental frequencies. Although simple, the harmonic termination control is very effective in enhancing the efficiency of RF transistors, especially when exploiting the Class J design space. The proposed dual-band matching network synthesis methodology was applied to design a dual-band power amplifier using a packaged 45 W gallium nitride (GaN) transistor. The power amplifier prototype maintained a peak power efficiency of about 68% at the two operating frequencies, namely 800 MHz and 1.9 GHz. In addition, a Volterra based digital predistortion technique has been successfully applied to linearize the PA response around the two operating frequencies. In fact, when driven with multi-carrier wideband code division multiple access (WCDMA) and long term evolution (LTE) signals, the linearized amplifier maintained an adjacent channel power ratio (ACPR) of about 50 dBc and 46 dBc, respectively.

A microwave module supplying up-to 1 Watt output power at 2.4-2.5 GHz frequency band was investigated. First the module was operated at low power levels and the output power was predicted using the small signal S-parameters of the module. A method was developed to obtain its large signal model using Advanced Design System (ADS) simulator&rsquo / s nonlinear analyses facilities. Later using the large signal model of the module simulations carried out to obtain larger powers up-to 1 Watt. The implementation of the module was performed using the SMD components on a microstrip substrate and the characteristics of the module were compared to the ones obtained using simulation results.

Technological developments in the area of high power LED light sources have enabled their utilization in general illumination applications. Along with this advancement comes the need for progressive thermal management strategies in order to ensure device performance and reliability. Minimizing an LED's junction temperature is done by minimizing the total system's thermal resistance. For actively cooled systems, this may essentially be achieved by simultaneously engineering the conduction through the heat sink and creating a well-designed flow pattern over suitable convective surface area. While such systems are routinely used in cooling microelectronics, their use in LED lighting systems encounter additional constraints which must be accounted for in the design of the cooling system. These are typically driven by the size, shape, and building codes involved with the lighting industry, and thus influence the design of drop-in replacement LED fixtures. Employing LED systems for customary down-lighting applications may require shrouded radial fin heat sinks to increase the heat transfer while reducing the space requirement for active cooling. Most lighting is already in some form of housing, and the ability to concurrently optimize these housings for thermal and optical performance could accelerate the widespread implementation of cost-efficient, environmentally-friendly solid-state lighting. In response, this research investigated the use of conical, cylindrical, square, and pyramidal shrouds with pin/radial fin heat sink designs for the thermal management of high power LED sources. Numerical simulations using FLUENT were executed in order to account for details of the air flow, pressure drop, and pumping power, as well as the heat transfer and temperature distributions throughout the system. The LEDs were modeled as a distributed heat source of 25 - 75 W on a central portion of the various heat sinks. Combinations of device junction temperature and pumping power were used to assess the performance of shrouded heat sink designs for their use in air-cooled, down-lighting LED fixtures.

Thesis (M.S. in Applied Physics)--Naval Postgraduate School, June 2005. / Thesis Advisor(s): William B. Colson. Includes bibliographical references (p. 79-81). Also available online.

Thesis (Ph. D.)--University of Texas at Austin, 2007. / Vita. Includes bibliographical references.

Thesis (Ph. D.)--University of Texas at Austin, 2007. / Vita. Includes bibliographical references.

Thesis (Ph. D.)--University of Texas at Austin, 2006. / Vita. Includes bibliographical references.

Les accélération de faisceaux de forte puissance et le bombardement d'une cible de spallation permettent d'atteindre les flux requis pour application aux réacteurs sous-critiques pilotés par un accélérateur de particules (ADSR).Parmi les methodes d'acceleration, on trouve les accélérateurs circulaires, en particuler les FFAG et les cyclotrons. Le travail de cette these porte sur l'évaluation de ces methodes pour la production de faisceaux de protons de classe multi-mégawatts, y compris la participation à des collaborations d'études internationales de design. Une evaluation de la methode ADSR comme candidat potential pour resoudre le probleme des dechets nucleaires est finalement etablie. / High power proton accelerators allow providing, by spallation reaction, the neutron fluxes necessary in thesynthesis of fissile material, starting from Uranium 238 or Thorium 232. This is the basis of the concept ofsub-critical operation of a reactor, for energy production or nuclear waste transmutation, with the objective ofachieving cleaner, safer and more efficient process than today’s technologies allow.Designing, building and operating a proton accelerator in the 500-1000 MeV energy range, CW regime,MW power class still remains a challenge nowadays. There is a limited number of installations at presentachieving beam characteristics in that class, e.g., PSI in Villigen, 590 MeV CW beam from a cyclotron, SNS inOakland, 1 GeV pulsed beam from a linear accelerator, in addition to projects as the ESS in Europe, a 5 MWbeam from a linear accelerator.Furthermore, coupling an accelerator to a sub-critical nuclear reactor is a challenging proposition: some ofthe key issues/requirements are the design of a spallation target to withstand high power densities as well asensure the safety of the installation.These two domains are the grounds of the PhD work: the focus is on the high power ring methods inthe frame of the KURRI FFAG collaboration in Japan: upgrade of the installation towards high intensityis crucial to demonstrate the high beam power capability of FFAG. Thus, modeling of the beam dynamicsand benchmarking of different codes was undertaken to validate the simulation results. Experimental resultsrevealed some major losses that need to be understood and eventually overcome.By developing analytical models that account for the field defects, one identified major sources of imperfectionin the design of scaling FFAG that explain the important tune variations resulting in the crossing of severalbetatron resonances. A new formula is derived to compute the tunes and properties established that characterizethe effect of the field imperfections on the transverse beam dynamics. The results obtained allow to developa correction scheme to minimize the tune variations of the FFAG. This is the cornerstone of a new fixed tunenon-scaling FFAG that represents a potential candidate for high power applications.As part of the developments towards high power at the KURRI FFAG, beam dynamics studies have toaccount for space charge effects. In that framework, models have been installed in the tracking code ZGOUBIto account for the self-interaction of the particles in the accelerator. Application to the FFAG studies is shown.Finally, one focused on the ADSR concept as a candidate to solve the problem of nuclear waste. In orderto establish the accelerator requirements, one compared the performance of ADSR with other conventionalcritical reactors by means of the levelized cost of energy. A general comparison between the different acceleratortechnologies that can satisfy these requirements is finally presented.In summary, the main drawback of the ADSR technology is the high Levelized Cost Of Energy comparedto other advanced reactor concepts that do not employ an accelerator. Nowadays, this is a show-stopper forany industrial application aiming at producing energy (without dealing with the waste problem). Besides, thereactor is not intrinsically safer than critical reactor concepts, given the complexity of managing the targetinterface between the accelerator and the reactor core.

Ce travail de thèse, réalisé au sein du CEA-LETI, consiste en la décorrélation des différents mécanismes ayant lieu dans une LED à base de GaN par voie de simulation numérique et de caractérisation expérimentale. Dans les chapitres 1 et 2, les théories des différents mécanismes présents dans une diode/LED sont décrites. Dans le chapitre 3, la simulation numérique d’une structure LED VTF ("vertical thin film" en langue anglaise) détermine les mécanismes prédominants dans les différentes gammes de tension. Une étude paramétrique s’ensuit afin d’évaluer les interactions entre les mécanismes.Dans le chapitre 4, les simulations sont effectuées en ajoutant un modèle de dépendance des mobilités des porteurs au champ électrique. En présence de ce modèle, les caractéristiques électro-optiques simulées tendent vers les caractéristiques réelles à haute tension.Les résultats des mesures de vitesse des porteurs dans le GaN-p sont également reportés dans le chapitre 4. En utilisant un motif d’échantillon spécifique et la méthode de mesure quatre points, les résultats suggèrent une tendance à saturer de la vitesse des porteurs, ou une tendance à décroître de la mobilité, à fort champ électrique. Ces résultats renforcent l’hypothèse de simulation utilisée dans le chapitre 4.Les simulations présentées dans les chapitres 3 et 4 permettent de proposer un schéma équivalent d’une LED en décorrélant les différents mécanismes et en ne gardant que les mécanismes dominants. Ce schéma équivalent permettrait, par exemple, à identifier les différents régimes dans une caractéristique électrique d’une LED réelle afin de procéder des améliorations du fonctionnement de la LED.Le chapitre 5 présente une étude par électroluminescence pulsée, une méthode de caractérisation fréquentielle, sur des LED commerciales. L’étude des temps de montée et de descente des signaux électro-optiques, ainsi que l’étude de durée de vie différentielle des porteurs de charge dans une LED, fourniraient des informations supplémentaires concernant l’injection des porteurs dans la LED. / This PhD. works, which was carried out inside CEA-LETI, aims to dissociate the various mechanisms occurring inside a GaN-based LED employing numerical simulation and experimental characterization. In the chapters 1 and 2, various mechanisms occurring inside a diode/LED are theoretically described. In chapter 3, through numerical simulation, the dominant mechanisms as well as their locations in a VTF ("vertical thin film") LED structure are determined for different voltage ranges. A parametric study follows to assess the interactions between the mechanisms.In chapter 4, the simulations are carried out with an additional field-dependent model for charge carrier mobility. With this model enabled, the simulated LED-electrical-and-optical characteristics approximate the real LED characteristics.Carrier-velocity characterization on p-type GaN, using a specific sample structure and the resistivity method, is also shown in chapter 4. It can be inferred from the results that under strong electric-fields, the carrier velocity might saturate, or the carrier mobility might decrease. These results strengthen the hypothesis used for the simulations in this chapter 4.The simulations introduced in the chapters 3 and 4 allow the proposition of an equivalent circuit for a GaN-based LED by dissociating the different mechanisms and retaining the dominant ones. This equivalent circuit could help, for instance, identify the different regimes in a real-LED electrical characteristics in order to improve the LED’s performance.Chapter 5 introduces pulsed electroluminescence, a frequential characterization method, on commercial LEDs. The studies of rise-time and fall-time of electro-optical signals, and the study of the differential lifetime of charge carriers in an LED would provide supplementary information concerning the carrier injection into the LED.

Le développement des générateurs de forte puissance pour les énergies marines renouvelables nécessite un effort important de recherche et de développement plus particulièrement centre sur les aspects mécaniques et thermiques. Généralement, la modélisation précise des machines électriques nécessite l'utilisation de la méthode des éléments finis. Cependant, cette méthode est gourmande en temps de calcul en particulier lors des premières étapes de conception. Le but de cette thèse est de fournir des modèles permettant l’analyse des générateurs synchrones à aimants permanents pour les énergies marines renouvelables avec un rapport temps / précision réduit. Deux approches de modélisation sont proposées dans ce travail: une approche à constantes localisées et une approche hybride basée sur le couplage direct entre un modèle analytique et un modèle à constantes localisées. Ces modèles sont utilisés pour l'étude des différents phénomènes physiques inhérents au fonctionnement des dispositifs électromagnétiques (magnétiques, thermiques et mécaniques). Les modèles magnétiques et thermiques à constantes localisées sont construits à partir d'une démarche dite générique, consistant en un maillage du domaine d'étude. Les modèles développés enrichiront les travaux en cours sur les modèles thermiques et mécaniques au GREAH afin d'aboutir à un code de calcul multi-physique par constantes localisées. / The development of high power generators for the marine renewable energies requires substantial research and development, focusing particularly on the mechanical and thermal aspects. Generally, accurate modelling of electrical machines requires the use of finite-element (fe) method. However, fe analysis is highly time consuming, especially at first design stages. The aim of this phd thesis is to propose a design approach of the permanent magnet synchronous generators for marine renewable energies applications offering a good computation time to precision ratio. Two design models are proposed in this work : lumped parameter approach and a hybrid approach based on the direct coupling of an analytical model and lumped parameter model. This modelling approach is used for the modelling of the three main physics to be considered in an electrical generator (magnetic model, thermal model and mechanical model). The lumped parameters approach consists, in this thesis, in the development of a generic magnetica and thermic lumped parameter network linked to the machine geometry. The models developed will regroup the ongoing developments of thermal and mechanical models at the GREAH lab in order to reach the completion of a mutliphysics lumped parameters computational code.