Ce travail porte sur les propriétés des nanocristaux de cellulose, leur modification de surface et le développement de matériaux avancés. Diverses approches sont utilisées sur ces substrats nanométriques visant à modifier leurs propriétés de surface et étendre leur utilisation dans des applications très sophistiquées, telles que la postsulfation et la désulfatation, le greffage et l'adsorption de polymères, l’oxydation sélective, le greffage moléculaire et l'inclusion "hôte-invité". Sur la base de modifications de surface, l'analyse des propriétés (pour différents taux de groupements sulfates) et divers nanomatériaux dérivés des nanocristaux de cellulose sont étudiés et préparés, notamment des nanocristaux sulfatés à différents taux, des nanocomposites extrudés, des éponges biocomposites et des hydrogels supramoléculaires. L'effet d’un gradient de groupements sulfates sur la chimie de surface, la morphologie et les propriétés physiques des nanocristaux de cellulose est discuté et notamment quatre modèles de section transversale sont comparés pour la détermination de la mesure du degré de substitution surfacique des nanocristaux de cellulose. Une stratégie nouvelle de protection impliquant une double couche polymère et la compatibilisation physique et/ou chimique des nanocristaux de cellulose est proposée afin de promouvoir à la fois la stabilité thermique des nanoparticules et la compatibilité des nanocristaux avec des matrices polymères non polaires au cours de la mise en forme par extrusion. En participant àla réticulation pour la construction de matériaux avancés, des nanocristaux de cellulose sélectivement oxydés (et de la cellulose microfibrillée oxydée pour comparaison) sont introduits dans de l'alginate pour développer des éponges biocomposites présentant une meilleure stabilité mécanique et une meilleure stabilité structurelle. Grâce à la conception intelligente par inclusion 'hôte-invité' in situ entre des nanocristaux de cellulose chimiquement modifiés et la cyclodextrine,deux polysaccharides hydrophiles sont combinés dans des hydrogels supramoléculaires pour l'administration de médicaments. En un mot, cette thèse contribue à l’avancée des nanocristaux de cellulose dans les domaines de l'analyse des propriétés et le développement des applications. / The present work focuses on the properties of cellulose nanocrystals, their surface modification and development of advanced materials. Diverse approaches are employed on these nanoscaled substrates aiming to modify their surface properties and extend their use in highly sophisticated applications, such as postsulfation and desulfation, polymer grafting and adsorption, selective oxidation, molecular grafting, and ‘host-guest' inclusion. On the basis of surface modifications, properties analysis (for different sulfate group contents) and various nanomaterials derived from cellulose nanocrystals are investigated and prepared, including gradient sulfated nanocrystals, extruded nanocomposites, biocomposite sponges, and supramolecular hydrogels. The effect of gradient degrees of sulfate groups on cellulose nanocrystals to surface chemistry, morphology and physical properties are discussed, particularly four cross-section models are compared for the determination of the surface degree of substitution on cellulose nanocrystals. A novel strategy involving a double-polymer-layer shield and physical and/or chemical compatibilization of cellulose nanocrystals is proposed, in order to realize both improvement of thermal stability and promotion of compatibility for nanocrystals with non-polar polymeric matrices during processing by melt-extrusion. With the idea of participating as crosslinking aid for the construction of advanced materials, selectively oxidized cellulose nanocrystals (with oxidized microfibrillated cellulose as comparison) are introduced in alginate for the development of biocomposite sponges with improved mechanical stability or structural stability. Through the smart design of in situ ‘host−guest' inclusion between chemically modified cellulose nanocrystals and cyclodextrin, two hydrophilic polysaccharides are combined in supramolecular hydrogels for use as drug delivery. In a word, this dissertation contributes to the advances of cellulose nanocrystals in the topics of property analysis and application development.

Cette thèse cible l'étude de la structure thermique de la croûte supérieure (<10km) dans les arcs magmatiques continentaux, et son influence sur l'enregistrement thermochronologique de leur exhumation et de leur évolution topographique. Nous portons notre regard sur deux chaînes de montagne appartenant aux Cordillères Américaines : Les Cascades Nord (USA) et la zone de faille Motagua (Guatemala). L'approche utilisée est axée sur l'utilisation de la thermochronologie (U-Th-Sm)/He sur apatite et zircon, couplée avec la modélisation numérique de la structure thermique de la croûte. Nous mettons en évidence la variabilité à la fois spatiale et temporelle du gradient géothermique, et attirons l'attention du lecteur sur l'importance de prendre en compte la multitude des processus géologiques perturbant la structure thermique dans les chaînes de type cordillère, c'est à dire formées lors de la subduction océanique sous un continent. / This thesis focuses on the influence of the dynamic thermal structure of the upper crust (<10km) on the thermochronologic record of the exhumational and topographic history of magmatic continental arcs. Two mountain belts from the American Cordillera are studied: the North Cascades (USA) and the Motagua fault zone (Guatemala). I use a combined approach coupling apatite and zircon (U-Th-Sm)/He thermochronology and thermo-kinematic numerical modelling. This study highlights the temporal and spatial variability of the geothermal gradient and the importance to take into account the different geological processes that perturb the thermal structure of Cordilleran-type mountain belts (i.e. mountain belts related to oceanic subduction underneath a continent).

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The variation on the richness species in regional scales has been debated a lot. Three of the hypotheses mot frequently discussed as determinants of regional variability in richness species are energy, contemporary climatic conditions and habitat heterogeneity. Studies at mesoscale are very important to discover different patterns that underlying the variability in richness species. As the bats are so common in whole world and ecologically diverse, these organisms are excellent environmental characterized. On this basis, the goals of this study were to link the bat species composition with the vegetation in forest-grassland transition area in the Southeastern of South America and to assess the space influence on the bat species richness by environmental variables. The study area includes the southeastern of South America. Species distribution data were obtained by museums, review and database on line .The distribution s estimate for each species was done on a map with 139 cells 1º latitude by 1º longitude. It was collected distribution data to 79 bat species. Rain Forest and Seasonal Forest had the largest species richness. The cluster analysis based on the cells revealed three major clusters: the Forest cluster (at north), the Araucaria cluster (at middle), and the Grassland cluster (at south). The Araucaria cluster was associated to the Grassland cluster. The cluster analysis based in the phytogeographic zones revealed four clusters, where the Araucaria Forest was associated with the forested regions. The contemporary climatic conditions hypothesis was the best explained the variability on the richness data, where the mean annual temperature the main predictor variable, followed by extent of relief and AET, respectively. With the whole variables, the OLS regression also indicated the mean annual temperature as the main predictor variable, followed by AET. These same variables also explained the variability in the species richness in Phyllostomidae, Molossidae and Vespertilionidae. The Araucaria Forest is a filter area to the dispersal of tropical species into subtropical and temperate areas of South America, because the hard winters in that area, what may be explained likely for the intolerance of some species in colder climates. / A variação da riqueza de espécies em escala regional tem sido intensivamente discutida. Três das principais hipóteses que buscam explicar esse padrão de variação são as hipóteses energética, climática e de heterogeneidade do habitat. Estudos em meso-escala são importantes para descobrir padrões diferentes que influenciam a variação na riqueza de espécies. Como os morcegos são muito abundantes globalmente e bem diversificados ecologicamente, esses organismos são ótimos caracterizadores ambientais. Com base nisso, os objetivos deste estudo foram relacionar a composição das espécies de morcegos com a vegetação em área de transição floresta-campo no sudeste da América do Sul e avaliar a influência do espaço sobre a riqueza de morcegos através de variáveis ambientais. A área de estudo compreende a porção sudeste da América do Sul. Os dados de distribuição das espécies foram coletados através de visita a museus, revisão bibliográfica e consulta a bancos de dados on line. Foi feita a estimativa de distribuição para cada espécie sobre um mapa contendo 139 quadrículas de 1º de latitude por 1º de longitude. Foram obtidos dados de distribuição para 79 espécies de morcegos. Áreas de Floresta Ombrófila Densa e de Floresta Estacional possuíram a maior riqueza de espécies. A análise de agrupamento com base nas quadrículas formou três grupos nítidos: grupo Florestal (ao norte), Araucária (intermediário) e Campestre (ao sul). O grupo Araucária ficou associado ao grupo Campestre. A análise de agrupamento com base nas regiões fitogeográficas formou quatro grupos nítidos, sendo que a Floresta de Araucária ficou associada às formações florestais. A hipótese climática foi a que melhor explicou a variação dos dados de riqueza, sendo a temperatura média anual a principal variável preditora, seguida pela amplitude da altitude e pela AET, respectivamente. Considerando todas as variáveis, a regressão OLS também apontou a temperatura média anual como a principal variável preditora, seguida pela AET. Essas variáveis também explicaram a variação na riqueza de espécies das famílias Phyllostomidae, Molossidae e Vespertilionidae. A Floresta de Araucária é uma área filtro para a dispersão de espécies tropicais para as regiões subtropicais e temperadas da América do Sul, devido aos invernos rigorosos presentes nessa área, o que pode ser explicado principalmente pela intolerância de algumas espécies a climas mais frios.

Dans le cadre de la vitrification des déchets nucléaires de haute activité à vie longue, un nouveau procédé a été mis en service à l’usine de La Hague en 2010 : le procédé creuset froid. Dans ce procédé, des gradients thermiques apparaissent au sein du bain de verre. Celui-ci forme une couche solide au contact de la paroi froide, appelée « auto-creuset ». Dans cette zone, le verre est soumis à des températures où il peut potentiellement cristalliser. L’objectif de ce travail était de déterminer la microstructure de cet auto-creuset en précisant les zones de cristallisation. Parallèlement, il s’agissait d’évaluer l’impact du gradient thermique sur la cristallisation des verres considérés. La cristallisation de deux verres d’intérêt nucléaire a donc été étudiée à l’aide d’une méthode basée sur l’analyse d’images MEB en conditions de traitements isotherme et sous gradient thermique. Les analyses en isotherme mettent en évidence la cristallisation de cristaux d’apatite (660°C-900°C) et de powellite (630°C-900°C) et permettent de quantifier cette cristallisation (vitesses de croissance et de nucléation, fraction cristallisée) qui reste très limitée (< 3%). La comparaison des résultats issus de ces deux types d’expérimentations montre que le gradient thermique n’a pas d’impact mesurable sur les cristallisations observées. Afin de compléter les analyses surfaciques de la cristallisation, des mesures par microtomographie in et ex situ ont été réalisées à l’ESRF sur la ligne ID19. Cette étude a permis de suivre la cristallisation d’apatites dans un verre simplifié et de confirmer la fiabilité de la méthode de quantification de la cristallisation basée sur l’analyse d’images 2D. / In the context of the vitrification of high level nuclear waste, a new industrial process has been launched in 2010 at the La Hague factory: The skull melting process. This setup applies thermal gradients to the melt, which leads to the formation of a solid layer of glass: the “self-crucible”. The question would be to know whether these thermal gradients have an impact or not on the crystallization behaviour of the considered glasses in the self crucible. In order to answer that question, the crystallization of two glass compositions of nuclear interest has been investigated with an image analysis based method in isothermal and thermal gradient heat treatments conditions. The isothermals experiments allow for the quantification (growth speed, nucleation, crystallized fraction) of the crystallization of apatites (660°C-900°C) and powellites (630°C- 900°C). The comparison of the results obtained through these two types of experimentations allows us to conclude that there is no impact of the thermal gradient on the crystallization of the studied glass compositions. In order to complete the image analysis study (based on surfaces), in and ex situ microtomography experiments have been performed at ESRF (Grenoble) on the ID10 beamline. This study allowed us to follow the crystallization of apatites in a simplified glass and to confirm the reliability of the image analysis method based on the analysis of surfaces.

O objetivo do presente estudo foi comparar o comportamento dos parâmetros mecânicos (Wext, Wint, Wtot, CP, FP), parâmetros energéticos (Pmet, C, Eff, Vótima) e do mecanismo pendular (R, Rint, %Cong) durante a caminhada com carga no plano (0%), nas inclinações (+7% e +15%) e em distintas velocidades de caminhada. A amostra foi composta por 10 homens jovens, saudáveis, fisicamente ativos e não adaptados ao transporte de carga em mochilas. Os sujeitos caminharam em uma esteira rolante durante 5 min, em cinco diferentes velocidades, sem e com carga (25% da MC) transportada em mochilas, e em três planos distintos de caminhada (0%, 7% e 15%). A análise de movimento 3D (quatro câmeras de vídeo) foi realizada simultaneamente à análise de VO2. Realizaram-se rotinas computacionais para o processamento de dados cinemáticos em Matlab®. Utilizou-se ANOVA de 3 fatores para medidas repetidas, com post hoc de Bonferroni (p < 0,05; SPSS 17.0). Os resultados dos parâmetros mecânicos indicam modificações devido à velocidade e ao plano de caminhada; a carga não modificou algumas das variáveis. Todas as variáveis mecânicas aumentaram com o incremento da velocidade, o Wint e a FP diminuíram a 7% e logo aumentaram a 15%, o Wext e Wtot aumentaram com a inclinação, e o CP diminuiu com o aumento da inclinação. A carga não afetou na maioria das situações o Wext e o Wtot, demonstrando que os parâmetros mecânicos são de modo geral, independentes da carga tanto no plano como nas inclinações. As variáveis energéticas da caminhada foram influenciadas pela velocidade, inclinação e a carga. A Pmet aumentou com o incremento da velocidade, da inclinação e da carga. O C diminuiu com o incremento da velocidade e logo aumentou, atingindo um mínimo nas velocidades intermediárias e, também aumentou com o incremento da inclinação e da carga. A Eff aumentou com a velocidade, diminuiu com o aumento da inclinação e a carga. A Vótima de caminhada foi reduzida com o incremento da inclinação. Constatou-se que o mecanismo pendular é modificado principalmente como decorrência da velocidade e da inclinação do terreno, e é independente da carga. O R e o Rint aumentam com o acréscimo da velocidade de caminhada, logo diminuem com o incremento da inclinação e ambos são independentes da carga. Conclui-se que as diferentes restrições impostas através da variação da carga e inclinações provocaram adaptações na mecânica e energética da locomoção humana, sustentando a Vótima e a reconversão das energias mecânicas (R) nas inclinações. Deste modo, ainda que em menor proporção, a estratégia de minimização de energia por via pendular ainda persiste nestas condições. / The purpose of the present study was to compare the behavior of the mechanical parameters (Wext, Wint, Wtot, SF, SL), the energetic parameters (Metabolic Power, C, Eff, optimal speed) and the pendular mechanism (R, Rint, %congruity) during walking with load on level (0%) and gradients (7% and 15%) and at different walking speeds. Ten young men, healthy, physically active and not adapted to walking loaded in backpacks participated in the study. The subjects walked in a treadmill for five minutes, under five different speeds, without and with load (25% of bM) carried in backpacks and in three different gradients of walking (0%, 7% e 15%). The analysis of the 3D movement was registered (four video cameras), as well as the VO2 analysis. Computational routines for the processing of kinematic data were done in Matlab®. The results were analyzed using repeated measures ANOVA (factors: speed, gradients, load) with the Bonferroni correction for post-hoc comparisons (p < 0.05) (SPSS 17.0). The results of the mechanical parameters indicate modifications due to speed and gradients of the walking; the load did not modify some of the variables. All of the mechanical variables increased with the raise in speed, the Wint and the SF decreased at 7% and right away increased at 15%, the Wext and Wtot increased with the gradient, and the SL decreased with the raising gradient. The load did not affect most of the situations, the Wext and Wtot, decreasing with the loaded situation, showing that the mechanical parameters are, in general, independent of the load on level and gradients. The energetic parameters of the walking were influenced by the speed, the gradient and the load. The metabolic power increased with the raise in speed, in gradient and in load. The cost of transport decreased with raise in speed and increased right away, influencing the minimum cost at intermediate speeds and it also increased with the raise of the slope and the load. The efficiency increased with speed and decreased with the raise of gradient and load. The optimal speed of walking was reduced with increasing of gradient. It was verified that the pendular mechanism is mainly modified as a consequence of speed and the gradient, and is independent of load. The R and Rint increase with the raise of speed, and decrease with the raise of gradient, also there are independent of the load. The conclusion is that the different restrictions imposed through the load variation and gradients cause adaptations in the mechanics and energetic of the human locomotion, sustaining the optimal speed and reconversion of the mechanical energies in gradients. In this way, but in a smaller proportion, the strategy of minimizing the energy through the pendular mechanism still persists in these conditions.

Cette thèse est principalement consacrée au développement de méthodes de sélection de modèles par maximum de vraisemblance pénalisée dans le cadre de données complexes. Un premier travail porte sur la sélection des modèles linéaires généralisés dans le cadre de données stratifiées, caractérisées par la mesure d’observations ainsi que de covariables au sein de différents groupes (ou strates). Le but de l’analyse est alors de déterminer quelles covariables influencent de façon globale (quelque soit la strate) les observations mais aussi d’évaluer l’hétérogénéité de cet effet à travers les strates.Nous nous intéressons par la suite à la sélection des modèles non linéaires à effets mixtes utilisés dans l’analyse de données longitudinales comme celles rencontrées en pharmacocinétique de population. Dans un premier travail, nous décrivons un algorithme de type SAEM au sein duquel la pénalité est prise en compte lors de l’étape M en résolvant un problème de régression pénalisé à chaque itération. Dans un second travail, en s’inspirant des algorithmes de type gradient proximaux, nous simplifions l’étape M de l’algorithme SAEM pénalisé précédemment décrit en ne réalisant qu’une itération gradient proximale à chaque itération. Cet algorithme, baptisé Stochastic Approximation Proximal Gradient algorithm (SAPG), correspond à un algorithme gradient proximal dans lequel le gradient de la vraisemblance est approché par une technique d’approximation stochastique.Pour finir, nous présentons deux travaux de modélisation statistique, réalisés au cours de cette thèse. / This thesis is mainly devoted to the development of penalized maximum likelihood methods for the study of complex data.A first work deals with the selection of generalized linear models in the framework of stratified data, characterized by the measurement of observations as well as covariates within different groups (or strata). The purpose of the analysis is then to determine which covariates influence in a global way (whatever the stratum) the observations but also to evaluate the heterogeneity of this effect across the strata.Secondly, we are interested in the selection of nonlinear mixed effects models used in the analysis of longitudinal data. In a first work, we describe a SAEM-type algorithm in which the penalty is taken into account during step M by solving a penalized regression problem at each iteration. In a second work, inspired by proximal gradient algorithms, we simplify the M step of the penalized SAEM algorithm previously described by performing only one proximal gradient iteration at each iteration. This algorithm, called Stochastic Approximation Proximal Gradient Algorithm (SAPG), corresponds to a proximal gradient algorithm in which the gradient of the likelihood is approximated by a stochastic approximation technique.Finally, we present two statistical modeling works realized during this thesis.

O objetivo do presente estudo foi comparar o comportamento dos parâmetros mecânicos (Wext, Wint, Wtot, CP, FP), parâmetros energéticos (Pmet, C, Eff, Vótima) e do mecanismo pendular (R, Rint, %Cong) durante a caminhada com carga no plano (0%), nas inclinações (+7% e +15%) e em distintas velocidades de caminhada. A amostra foi composta por 10 homens jovens, saudáveis, fisicamente ativos e não adaptados ao transporte de carga em mochilas. Os sujeitos caminharam em uma esteira rolante durante 5 min, em cinco diferentes velocidades, sem e com carga (25% da MC) transportada em mochilas, e em três planos distintos de caminhada (0%, 7% e 15%). A análise de movimento 3D (quatro câmeras de vídeo) foi realizada simultaneamente à análise de VO2. Realizaram-se rotinas computacionais para o processamento de dados cinemáticos em Matlab®. Utilizou-se ANOVA de 3 fatores para medidas repetidas, com post hoc de Bonferroni (p < 0,05; SPSS 17.0). Os resultados dos parâmetros mecânicos indicam modificações devido à velocidade e ao plano de caminhada; a carga não modificou algumas das variáveis. Todas as variáveis mecânicas aumentaram com o incremento da velocidade, o Wint e a FP diminuíram a 7% e logo aumentaram a 15%, o Wext e Wtot aumentaram com a inclinação, e o CP diminuiu com o aumento da inclinação. A carga não afetou na maioria das situações o Wext e o Wtot, demonstrando que os parâmetros mecânicos são de modo geral, independentes da carga tanto no plano como nas inclinações. As variáveis energéticas da caminhada foram influenciadas pela velocidade, inclinação e a carga. A Pmet aumentou com o incremento da velocidade, da inclinação e da carga. O C diminuiu com o incremento da velocidade e logo aumentou, atingindo um mínimo nas velocidades intermediárias e, também aumentou com o incremento da inclinação e da carga. A Eff aumentou com a velocidade, diminuiu com o aumento da inclinação e a carga. A Vótima de caminhada foi reduzida com o incremento da inclinação. Constatou-se que o mecanismo pendular é modificado principalmente como decorrência da velocidade e da inclinação do terreno, e é independente da carga. O R e o Rint aumentam com o acréscimo da velocidade de caminhada, logo diminuem com o incremento da inclinação e ambos são independentes da carga. Conclui-se que as diferentes restrições impostas através da variação da carga e inclinações provocaram adaptações na mecânica e energética da locomoção humana, sustentando a Vótima e a reconversão das energias mecânicas (R) nas inclinações. Deste modo, ainda que em menor proporção, a estratégia de minimização de energia por via pendular ainda persiste nestas condições. / The purpose of the present study was to compare the behavior of the mechanical parameters (Wext, Wint, Wtot, SF, SL), the energetic parameters (Metabolic Power, C, Eff, optimal speed) and the pendular mechanism (R, Rint, %congruity) during walking with load on level (0%) and gradients (7% and 15%) and at different walking speeds. Ten young men, healthy, physically active and not adapted to walking loaded in backpacks participated in the study. The subjects walked in a treadmill for five minutes, under five different speeds, without and with load (25% of bM) carried in backpacks and in three different gradients of walking (0%, 7% e 15%). The analysis of the 3D movement was registered (four video cameras), as well as the VO2 analysis. Computational routines for the processing of kinematic data were done in Matlab®. The results were analyzed using repeated measures ANOVA (factors: speed, gradients, load) with the Bonferroni correction for post-hoc comparisons (p < 0.05) (SPSS 17.0). The results of the mechanical parameters indicate modifications due to speed and gradients of the walking; the load did not modify some of the variables. All of the mechanical variables increased with the raise in speed, the Wint and the SF decreased at 7% and right away increased at 15%, the Wext and Wtot increased with the gradient, and the SL decreased with the raising gradient. The load did not affect most of the situations, the Wext and Wtot, decreasing with the loaded situation, showing that the mechanical parameters are, in general, independent of the load on level and gradients. The energetic parameters of the walking were influenced by the speed, the gradient and the load. The metabolic power increased with the raise in speed, in gradient and in load. The cost of transport decreased with raise in speed and increased right away, influencing the minimum cost at intermediate speeds and it also increased with the raise of the slope and the load. The efficiency increased with speed and decreased with the raise of gradient and load. The optimal speed of walking was reduced with increasing of gradient. It was verified that the pendular mechanism is mainly modified as a consequence of speed and the gradient, and is independent of load. The R and Rint increase with the raise of speed, and decrease with the raise of gradient, also there are independent of the load. The conclusion is that the different restrictions imposed through the load variation and gradients cause adaptations in the mechanics and energetic of the human locomotion, sustaining the optimal speed and reconversion of the mechanical energies in gradients. In this way, but in a smaller proportion, the strategy of minimizing the energy through the pendular mechanism still persists in these conditions.

Made available in DSpace on 2014-07-29T16:02:22Z (GMT). No. of bitstreams: 1 Dissertacao kelvin.pdf: 379268 bytes, checksum: 69875c577ac81dd2f77bb73f65c9f683 (MD5) Previous issue date: 2009-12-16 / The Convergence theory of gradient method and gradient projection method, for minimization of continuously differentiable generalized convex functions, that is, pseudoconvex functions and quasiconvex functions is studied in this work. We shall see that under certain conditions the gradient method, as well as gradient projection method, generate a convergent sequence and the limit point is a minimizing, whenever the function has minimizing and is pseudoconvex functions. If the objective function is quasiconvex then the generated sequence converges to a stationary point whenever that point exists. / Neste trabalho trataremos da convergência do método do gradiente para minimizar funções continuamente diferenciáveis e convexas-generalizadas, isto é, pseudo-convexas ou quase-convexas. Veremos que sob certas condições o método do gradiente, assim como o método do gradiente projetado, gera uma sequência que converge para minimizador quando existe um e a função objetivo é pseudo-convexa. Quando a função objetivo é quase-convexa a sequência gerada converge para um ponto estacionário do problema quando existe um tal ponto.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-03-07T14:25:21Z No. of bitstreams: 1 victorteixeirademelomayrink.pdf: 2587774 bytes, checksum: 1319cc37a15480796050b618b4d7e5f7 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-03-07T15:06:57Z (GMT) No. of bitstreams: 1 victorteixeirademelomayrink.pdf: 2587774 bytes, checksum: 1319cc37a15480796050b618b4d7e5f7 (MD5) / Made available in DSpace on 2017-03-07T15:06:57Z (GMT). No. of bitstreams: 1 victorteixeirademelomayrink.pdf: 2587774 bytes, checksum: 1319cc37a15480796050b618b4d7e5f7 (MD5) Previous issue date: 2016-08-31 / FAPEMIG - Fundação de Amparo à Pesquisa do Estado de Minas Gerais / O armazenamento de energia elétrica em larga escala ainda não é viável devido a restrições técnicas e econômicas. Portanto, toda energia consumida deve ser produzida instantaneamente; não é possível armazenar o excesso de produção, ou tampouco cobrir eventuais faltas de oferta com estoques de segurança, mesmo que por um curto período de tempo. Consequentemente, um dos principais desafios do planejamento energético consiste em realizar previsões acuradas para as demandas futuras. Neste trabalho, apresentamos um modelo de previsão para o consumo de energia elétrica a curto prazo. A metodologia utilizada compreende a construção de um comitê de previsão, por meio da aplicação do algoritmo Gradient Boosting em combinação com modelos de árvores de decisão e a técnica de amortecimento exponencial. Esta estratégia compreende um método de aprendizado supervisionado que ajusta o modelo de previsão com base em dados históricos do consumo de energia, das temperaturas registradas e de variáveis de calendário. Os modelos propostos foram testados em duas bases de dados distintas e demonstraram um ótimo desempenho quando comparados com resultados publicados em outros trabalhos recentes. / The storage of electrical energy is still not feasible on a large scale due to technical and economic issues. Therefore, all energy to be consumed must be produced instantly; it is not possible to store the production leftover, or either to cover any supply shortages with safety stocks, even for a short period of time. Thus, one of the main challenges of energy planning consists in computing accurate forecasts for the future demand. In this paper, we present a model for short-term load forecasting. The methodology consists in composing a prediction comitee by applying the Gradient Boosting algorithm in combination with decision tree models and the exponential smoothing technique. This strategy comprises a supervised learning method that adjusts the forecasting model based on historical energy consumption data, the recorded temperatures and calendar variables. The proposed models were tested in two di erent datasets and showed a good performance when compared with results published in recent papers.

Le poling thermique est une technique consistant à appliquer un fort champ électrique (DC) à un substrat de verre chauffé. Après traitement, un champ électrique est figé au sein de la matrice vitreuse, brisant sa centrosymmétrie. La présence de ce champ permet d’accéder à des propriétés d’optique nonlinéaire du second ordre, habituellement interdite dans un milieu centrosymmétrique tel que le verre. En plus des propriétés d’optique nonlinéaire, la présence du champ électrique a été associée à des modifications structurelles et compositionnelles mais également à des changements de propriétés de surface. Notre objectif a été d’utiliser cette technique pour modifier les propriétés de réactivité de surface et optique de verres d’oxyde (borosilicate et borophosphate de niobium (BPN)) et de verres de chalcogénures. Après poling, les modifications structurelles ont été caractérisée par spectroscopie vibrationnelle Raman et infrarouge. L’intensité et la localisation du champ électrique ont été caractérisées par des techniques de génération de seconde harmonique (SHG) : une analyse quantitative avec les franges de Maker et une d’imagerie μSHG. Le traitement a permis d’augmenter localement la réactivité de surface du verre borosilicate. Dans les verres BPN et chalcogénures, le traitement a permis de contrôler les propriétés optiques à la fois linéaire et nonlinéaire à l’échelle micrométrique. Ces résultats permettent d’envisager l’utilisation du poling thermique pour des applications en photonique intégrée. / Thermal poling is a technique which consists in the application of a strong DC electric field to a heated glass substrate. Following the treatment, a static electric field is frozen inside the glass matrix, effectively breaking its centrosymmetry. Presence of the electric field allows for second order non-linear optical properties usually forbidden in centrosymmetric medium such as glasses. In addition to nonlinear optical properties, the presence of the electric field has been associated with structural/compositional modifications as well as surface property changes. Our objective was to use this technique to tailor surface reactivity and optical properties in oxide (borosilicate and niobium borophosphate) and chalcogenide glasses. After poling, structural modifications were investigated using Raman and infrared spectroscopy. Strength and localization of the electric field were characterized by Second Harmonic Generation (SHG) techniques: quantitative Maker fringes analysis and μSHG imaging. The treatment has successfully allowed to locally enhanced the surface reactivity of a borosilicate glass. In niobium borophosphate and chalcogenide glasses, the treatment has permitted to control optical properties both linearly and non-linearly at the micrometric scale. These results show that thermal poling could be used to create functional devices for applications in integrated photonics.