No description available.

In this thesis, the design of passive networks with the aid of full-wave simulation software and geometry-based models of lumped elements is investigated. This is done by examining the results of a number of simulation examples, as well as measured data from manufactured designs to compare against simulated equivalents. One such example is a chip antenna evaluation board design, in which the PCB, antenna, matching components and connector are all modeled. When measured, the simulation agreed with the board’s best matched frequency of 5.5 GHz to within 20 MHz. In another, a new antenna layout is generated from an existing evaluation design which, produced a match of about -15 dB at the design frequency with a similar bandwidth to that shown on the antenna datasheet on the first attempt at manufacture. Additionally, a statistical experiment was conducted in order to provide insight into the phenomenon of coupling between lumped components, and to define clearly when it starts to become an important effect to consider. For both chip capacitors and inductors, a behavioral model of how much crosstalk is present in a prospective circuit was developed which takes into account angle and distance between components, as well as case size. Finally, a simple discrete gradient descent was implemented in a commercial full-wave simulation software in order to assist in the refinement of designs containing 3-D geometry-defined component models.

Au cours du traitement thermique rapide (RTP) des cellules photovoltaïques à couches minces, un suivi du profil de température souhaité et une homogénéité de la chauffe substrat doivent être assurés. Le but de cette thèse est de proposer une méthodologie de conception d'un four RTP permettant d'atteindre la qualité du cycle de la chauffe souhaitée.Une modélisation thermique est réalisée en se basant sur la méthode de réseaux de composants afin de prédire le comportement thermique dynamique du four. L'approximation des flux plans et l'approximation des couches minces semi-transparentes sont utilisées pour le calcul des facteurs d'échanges directs. L'algorithme des revêtements est appliqué pour en déduire les facteurs de transfert. Le modèle thermique développé est validé expérimentalement sur un four de petites dimensions. Une méthodologie de conception du four RTP est proposée en tenant compte de l'aspect dynamique des conditions thermiques du four. Une optimisation par algorithme génétique est effectuée pour trouver l'emplacement des émetteurs. Pour chacune des configurations testées, la distribution de la puissance aux émetteurs à fournir à chaque instant est optimisée par programmation dynamique. Finalement, cette méthodologie est appliquée pour la conception d'un four RTP pour le traitement de cellules photovoltaïques à couches minces de 30 × 60 cm2. Les résultats des essais confirment la validité de la méthodologie proposée. / During the rapid thermal processing (RTP) of thin film photovoltaic cells, the temperature of the latter has to follow a preset time evolution profile, while keeping spatial uniformity of the wafer. The aim of this study is to propose a design methodology of RTP furnace in order to obtain the quality of the required heating cycle.A thermal modeling is performed based on the component interaction network approach to predict the thermal behavior of the furnace. Flux plane approximation and semi-transparent thin layer approximation are used to calculate the direct exchange factor. The plating algorithm is then applied to calculate the transfer factor. The thermal model developed is validated experimentally on a furnace of small dimensions. A methodology to design a RTP furnace is proposed taking into account the dynamic aspect of the thermal conditions of the furnace. An optimization using the genetic algorithm is performed in order to find emitter dispositions. For each tested configuration, the optimal input power distribution over the emitters at each time step is found by using real time dynamic programming. Finally, the methodology is applied for the design of RTP furnace for the heat treatment of thin film photovoltaic cells of 30 × 60 cm2. Test results confirm the validity of the methodology proposed.