Na ocorrência de uma ou de múltiplas faltas permanentes em sistemas de distribuição de energia elétrica é inevitável a ausência de fornecimento à algumas cargas. Contudo, devido à presença de chaves seccionadoras, após a localização e isolação do trecho sob falta, é possível restaurar o abastecimento de energia a estas cargas fora de serviço. Para tal é necessária a obtenção de um plano de restabelecimento adequado, no menor intervalo de tempo possível, capaz de informar as chaves que devem ser operadas a fim de reconectar estas cargas à rede e sem sobrecarregar nenhum equipamento. Neste sentido, têm sido propostas metodologias para auxiliar as atividades dos operadores de sistemas de distribuição por meio do fornecimento de planos de restabelecimento de energia em situações de contingência. Todavia, a aplicação da maioria destas técnicas restringe-se à redes pequenas, quando comparadas às redes reais que possuem milhares de barras e chaves. Em outras metodologias, na tentativa de superar esta limitação, são feitas simplificações na representação da rede, desconsiderando alguns dos seus elementos. Nestes casos, a solução fornecida para a rede simplificada pode não ter o mesmo desempenho na rede em operação, o que compromete a confiabilidade dessas metodologias. Em outros métodos, ainda, a rede é representada com todas as suas barras e chaves, no entanto, deixam de considerar a presença de consumidores especiais, que exigem prioridade de fornecimento, bem como de chaves existentes na rede que podem ser operadas remotamente. Face ao exposto, propõe-se uma metodologia para determinação, em tempo real, de planos de restabelecimento de energia elétrica em sistemas de distribuição de grande porte em situações de uma ou de múltiplas faltas. Para lidar com os múltiplos objetivos deste problema, sendo alguns dele conflitantes entre si, esta será baseada em Algoritmos Evolutivos Multi-Objetivo, enquanto que a representação computacional sem simplificações da rede será proporcionada por uma codificação de dados apoiada na teoria de grafos e denominada Representação Nó-Profundidade. A fim de diminuir o tempo e o custo de implementação dos planos obtidos, será considerada e dado prioridade de manobra às chaves controladas remotamente, que podem ser alteradas do centro de operações. Será dado prioridade também ao fornecimento de energia de consumidores especiais. Por fim, para cada plano de restabelecimento fornecido será definida uma sequência factível de chaves que, quando alteradas, reconectarão as cargas sãs fora de serviço e eliminarão sobrecargas ou perfis de tensão inadequados. Para validar a metodologia proposta serão realizadas simulações computacionais no sistema de distribuição real da cidade de São Carlos-SP, em operação no ano de 1994, e nas suas versões duplicada, quadruplicada e octuplicada. / The occurrence of one or multiple permanent faults in electric power distribution systems is inevitable the privation of the service to some loads. However, due to the presence of switches, after the location and isolation of the faulted section, you can restore the power supply to these out of service loads. To this it is necessary to obtain, as soon as possible, a suitable restoration plan. It must be able to inform the switches that must be operated in order to reconnect these loads without overloading on any network or equipment. In this sense, it has been proposed methods to assist the operators\' activities of the of distribution systems by providing plans to restore the service in contingency situations. However, the application of the most of these techniques is limited to small networks, when they are compared to real networks that have thousands of bars and keys. In other methodologies, in order to overcome this limitation, simplications are made in the representation of the network, ignoring some of its elements. In these cases, the solution provided for a simplied network may not have the same performance on the network in operation, which aects the reliability of these methodologies. In other methods, the network is represented with all of its bars and switches. However, these methodologies do not consider the presence of the special consumers that require service priority neither the switches that can be operated remotely. Given the above, we propose a methodology for determining, in real time, plans for service restoration in large-scale electric distribution systems in situations of one or multiple faults. To deal with the multiple objectives of this problem, some of them con icting, the proposed method will be based on Multi-Objective Evolutionary Algorithms. Moreover, the computational representation without simplications of the network will be provided by a data encoding based in graph theory and called Node-Depth Encoding. In order to reduce the time and cost of the implementation of the plans obtained, it will be considered and given priority to maneuvers in remotely controlled switches, which can be changed from the operation center. Priority will be given also to the power supply of special consumers. Finally, for each restoration plan provided will be gotten a feasible switching sequence that, when implemented, will reconnect the out of service loads and eliminate overloads or inadequate voltage proles. To validate the proposed methodology it will be performed computer simulations in the real distribution system of the São Carlos-SP city, in operation in 1994, and their doubled, quadruplicate and octuplicate versions.

One of the primary targets of current and especially future cosmological observations are light thermal relics of the hot big bang. Within the Standard Model of particle physics, an important thermal relic are cosmic neutrinos, while many interesting extensions of the Standard Model predict new light particles which are even more weakly coupled to ordinary matter and therefore hard to detect in terrestrial experiments. On the other hand, these elusive particles may be produced efficiently in the early universe and their gravitational influence could be detectable in cosmological observables. In this thesis, we describe how measurements of the cosmic microwave background (CMB) and the large-scale structure (LSS) of the universe can shed new light on the properties of neutrinos and on the possible existence of other light relics. These cosmological observations are remarkably sensitive to the amount of radiation in the early universe, partly because free-streaming species such as neutrinos imprint a small phase shift in the baryon acoustic oscillations (BAO) which we study in detail in the CMB and LSS power spectra. Building on this analytic understanding, we provide further evidence for the cosmic neutrino background by independently confirming its free-streaming nature in different, currently available datasets. In particular, we propose and establish a new analysis of the BAO spectrum beyond its use as a standard ruler, resulting in the first measurement of this imprint of neutrinos in the clustering of galaxies. Future cosmological surveys, such as the next generation of CMB experiments (CMB-S4), have the potential to measure the energy density of relativistic species at the sub-percent level and will therefore be capable of probing physics beyond the Standard Model. We demonstrate how this improvement in sensitivity can indeed be achieved and present an observational target which would allow the detection of any extra light particle that has ever been in thermal equilibrium. Interestingly, even the absence of a detection would result in new insights by providing constraints on the couplings to the Standard Model. As an example, we show that existing bounds on additional scalar particles, such as axions, may be surpassed by orders of magnitude.

The loss of efficiency and performance of bioprocesses on scale-up is well known, but not fully understood. This work addresses this problem, by studying the effect of some fermentation gradients (pH, glucose and oxygen) at a larger scale in a bench-scale two compartment reactor (PFR + STR) using the cadaverine-producing recombinant bacterium, Corynebacterium glutamicum DM1945 Δact3 Ptuf-ldcC_OPT. The initial scale down strategy increased the magnitude of these gradients by only increasing the mean cell residence time in the plug flow reactor (τ_PFR). The cell growth and product related rate constants were compared as the τ_PFR was increased; differences were significant in some cases, but only up to 2 min residence time. For example, losses in cadaverine productivity when compared to the control fed-batch fermentation on average for the τ_PFR of 1 min, 2 min and 5 min were 25 %, 42 % and 46 % respectively. This indicated that the increasing the τ_PFR alone does not necessarily increase the magnitude of fermentation gradients. The new scale-down strategy developed here, increased the magnitude of fermentation gradients by not only increasing the τ_PFR, but also considering the mean frequency at which the bacterial cells entered the PFR section (f_m). The f_m was kept constant by reducing the broth volume in the STR. Hence, the bacterial cells also spent shorter times in the well mixed STR, as the τ_PFR was increased (hypothesised as giving the bacterial cells less time to recover the non-ideal PFR section of the SDR). On adoption of this strategy cadaverine productivity decreases for the τ_PFR of 1 min, 2 min and 5 min were 25 %, 32 % and 53 % respectively. Thus, highlighting that loss in performance is most likely to occur as the magnitude of heterogeneity within the fermentation environment increases. However, Corynebacterium glutamicum DM1945 Δact3 Ptuf-ldcC_OPT did show some resilience in its biomass productivity. It was only marginally affected in the harshest of conditions simulated here.

La présente thèse vise à caractériser le comportement mécanique à la ruine de poutres métalliques avec ouvertures, de formes et de tailles variables, renforcées ou non par raidisseurs. L’étude s’intéresse à des poutres de dimensions réalistes et s’appuie sur une combinaison de différentes approches : essais, modèles éléments finis et modèles analytiques. Les raidisseurs sont généralement utilisés pour corriger certaines faiblesses au niveau du comportement mécanique autour des ouvertures. Autour des ouvertures rectangulaires, ces raidisseurs, disposés verticalement ou horizontalement, ont fait l’objet de peu d’études scientifiques publiées. Généralement, le comportement des poutres à ouvertures d’âmes est considéré comme étant similaire à celui d’une poutre échelle (ou poutre Vierendeel) chargée aux noeuds. Cette hypothèse permet de considérer que les membrures de l’ouverture sont soumises à un moment fléchissant local bi-triangulaire sur la longueur de l’ouverture. Cependant, la distribution du moment fléchissant, et de ses contraintes axiales, le long d’une membrure peut s’avérer être plus complexe et dépend de la forme de l’ouverture (rectangulaire, hexagonale, circulaire, sinusoïdale ou quelconque), de la position de l’ouverture le long de la poutre et enfin, du type de chargement (concentré ou uniformément réparti). Plusieurs études numériques et expérimentales ont déjà été menées sur les poutres à ouvertures d’âmes et ont permis de développer différentes approches analytiques en vue de décrire leurs comportements. Cependant, ces modèles montrent certaines imprécisions ou insuffisances pour différents types et dimensions d’ouvertures notamment les ouvertures rectangulaires. Les travaux présentés dans cette thèse font le point sur les principaux modèles existants de calcul de résistance des poutres avec ouvertures d’âmes. Ils présentent notamment leurs avantages et limites. Pour conforter les descriptions basées sur les analyses bibliographiques, une campagne expérimentale est réalisée pour valider un modèle éléments finis développé dans le cadre de l’étude. Les essais s’appuient sur des tronçons de poutres de dimensions réalistes avec différentes configurations d’ouvertures isolées renforcées ou non par des raidisseurs. Les mesures autour de l’ouverture utilisent des jauges de déformation pour analyser le profil des contraintes et son évolution, au cours du chargement, dans différentes sections critiques. Les essais sont menés jusqu’à la ruine qui s’est développée généralement par une combinaison de comportement élasto-plastique et d’instabilité locale autour des ouvertures. Certaines ruines se sont produites par rupture en traction-cisaillement de l’acier dans des zones critiques. Les analyses numériques et expérimentales ont permis d’étudier l’influence de différentes conditions sur la distribution des contraintes, notamment axiales, au sein d’une membrure d’ouverture, en vue d’évaluer son potentiel de voilement local qui reste un phénomène complexe à appréhender. Elles ont permis aussi de comprendre le fonctionnement mécanique des ouvertures renforcées par des raidisseurs de différentes dimensions et positions. L’étude paramétrique menée, en utilisant le modèle éléments finis validé par essai, a permis d’évaluer la pertinence des hypothèses retenues dans les approches analytiques existantes. Ces analyses ont permis aussi de proposer un modèle analytique tenant compte du comportement d’ouvertures isolées avec ou sans raidisseurs. / The present thesis aims to characterize the mechanical behavior to failure of steel beams with openings, with variable shapes and sizes, reinforced or not by stiffeners. The study considers beams of realistic dimensions and combines different approaches: tests, finite element models and analytical models. The stiffeners are generally used to improve some weaknesses in the mechanical behavior around the openings. Around rectangular openings, these stiffeners, in vertical or horizontal arrangements, have been the subject of limited number of scientific publications. In general, the behavior of beams with web openings is considered similar to that of Vierendeel beams with loads applied at the nodes. This hypothesis allows to consider that each frame around the opening is characterized by a bi-triangular local bending moment along the frame elements. In fact, the bending moment distribution, and their resultant axial stresses, along a frame can be more complex and depends on the shape of the opening (rectangular, hexagonal, circular, sinusoidal or any), the position of the opening along the beam and finally, the type of loading (concentrated or uniformly distributed). Several numerical and experimental studies have been conducted on the beams with web openings and different analytical approaches have been developed to describe and predict the behavior of these beams. However, these models show some inaccuracies or inadequacies depending on the types and dimensions of openings including rectangular openings. The work presented in this thesis examines the main existing models predicting the resistance for beams with web openings. It shows their advantages and limits based on the existing but limited results. To obtain additional and specific results to be used in the comparisons and to validate a finite element model developed within the framework of the study, an experimental campaign is carried out. The tests are focused on beams of real dimensions with different configurations of isolated openings reinforced or not by stiffeners. The measurements around the opening use strain gauges to analyze the stress profile and its evolution, during loading, in different critical sections. The tests are carried out until failure generally characterized by a combination of elastic-plastic behavior and local instability around openings. Some final failures occurred by tensile-shear fracture of steel in some critical zones. Numerical and experimental analyzes are performed to study the influence of various conditions on the stresses distributions, in particular axial stresses, within the frames around the openings. The stress distribution is mainly observed to evaluate the local buckling potential that remains a complex phenomenon difficult to predict. The results are also analyzed to understand the mechanical behavior of the openings reinforced by stiffeners of different dimensions and positions. The parametric study conducted using the validated finite element model allows evaluating the relevance of the assumptions considered in the existing analytical approaches. These analyzes made it possible to propose an analytical model taking into-account the behavior of isolated openings with or without stiffeners.

Submitted by Silvana Teresinha Dornelles Studzinski (sstudzinski) on 2017-04-19T13:31:55Z No. of bitstreams: 1 JULIANA FATIMA SERRAGLIO PASINI_.pdf: 2321629 bytes, checksum: 991ec935f9ab785e766e00af290db5aa (MD5) / Made available in DSpace on 2017-04-19T13:31:56Z (GMT). No. of bitstreams: 1 JULIANA FATIMA SERRAGLIO PASINI_.pdf: 2321629 bytes, checksum: 991ec935f9ab785e766e00af290db5aa (MD5) Previous issue date: 2016-12-19 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O foco dessa pesquisa é o modo como as avaliações em larga escala repercutem no contexto da prática, tomando como espaço empírico cinco municípios de pequeno porte (com até 10 mil habitantes) do estado do Paraná. A abordagem teórico-metodológica considera que as políticas se desenvolvem em contextos de disputas contemplando arenas, lugares e grupos de interesses. Nessa tese o contexto de influência caracteriza-se pelo histórico da implementação das políticas de avaliação em âmbito federal e estadual e o contexto da prática, pela pesquisa com foco em ações desenvolvidas em âmbito escolar. A investigação teve como objetivo geral analisar a relação entre as políticas de avaliação em larga escala e as desenvolvidas em municípios de pequeno porte, a fim de identificar como repercutem no contexto da prática escolar, e, como objetivos específicos: (a) levantar os indicadores sociais e educacionais de cada município da pesquisa, no período de 2005 a 2013; (b) descrever as estratégias da gestão escolar e as implicações para o contexto da prática; (c) identificar as políticas educacionais desenvolvidas em âmbito municipal, a fim de promover melhorias no IDEB no período de 2005 a 2013; e (d) compreender como as políticas em nível federal e estadual repercutem nas políticas municipais. A coleta de dados desenvolveu-se com foco em ações desenvolvidas em âmbito escolar, mediante entrevistas (25) com diretores, coordenadores pedagógicos e professores, de escolas que participaram de, pelo menos, um dos ciclos da Prova Brasil, no período de 2005 a 2013. Utilizou-se o NVivo 11 para análise e organização do material obtido nas entrevistas. Como resultados, a pesquisa aponta fragilidades nas estatísticas quando apresentadas como fotografia da realidade educacional, o que acontece quando as redes e a sociedade tratam o IDEB como único ou principal instrumento para indicar a qualidade da educação, desconsiderando as especificidades de cada local e homogeneizando as características socioeconômicas, culturais e educacionais. Também revela que, mesmo em municípios de pequeno porte, a performatividade e o gerencialismo permeiam o contexto da prática, envolvendo ações relacionadas ao atendimento das especificidades das avaliações em larga escala. Entretanto, há resistência dos profissionais para que esses princípios não sejam determinantes da prática pedagógica escolar. Verifica-se grande preocupação com a aprendizagem dos alunos, desenvolvimento de projetos e atividades de contra turno, embora os resultados das avaliações em larga escala não sejam ignorados. / This research focuses on the way large-scale evaluations reverberate in the context of practice, taking as empirical space five small cities (with up to 10 thousand inhabitants) of the state of Paraná. Its theoretical-methodological approach considers that policies develop in contexts of dispute, contemplating arenas, places and groups of interest. In this thesis, the context of influence is characterized by the history of implementation of evaluation policies in federal and state spheres and by the context of practice; it focuses on actions developed in school settings. The main goal of this thesis was to analyze the relation between large-scale evaluation policies and the policies developed in small cities, in order to verify how they reverberate in the context of school practice; the specific objectives were: (a) to verify the social and educational indicators of each city in which the research was carried, in the period between 2005 and 2013; (b) to describe the school management strategies and its implications for the context of practice; (c) to identify the educational policies developed in municipal level, in order to promote improvements in IDEB in the period between 2005 and 2013; and (d) to understand how policies in federal and state levels reverberate in municipal policies. The collection of data focused in actions developed in school environments, by means of 25 interviews with directors, pedagogical coordinators and teachers from schools that participated in, at least, one of the cycles of Prova Brasil, in the period between 2005 and 2013. NVivo 11 was used to analyze and organize the material obtained during these interviews. As results, the research points the fragilities of the statistics when they are presented as a reflection of educational reality, which happens when educational networks and society take IDEB as the only or main tool for indicating education quality, disregarding the specificities of each place and homogenizing social and educational characteristics. It also shows that, even in small cities, performativity and managerialism permeate the context of practice, involving actions related to the meeting of large-scale evaluations' specificities. However, there is opposition from professionals so that these principles are not decisive in school pedagogical practice. Great concern with the learning of students is verified, as well as the development of projects and after-school activities, although the large-scale evaluation results are not ignored.

Quatre événements El Niño extrêmes ont eu lieu durant les cinq dernières décennies (1972/1973, 1982/1983, 1997/1998 et 2015/2016) et étaient caractérises comme forts dans la région Niño 3.4. Ils présentent des différences significatives dans leur évolution qui induisent des anomalies distinctes de précipitations le long du versant Pacifique Péruvien illustrant la non-linéarité de la téléconnexion ENSO sur les précipitations dans cette région. Les pluies extrêmes ont un impact néfaste sur la population et les secteurs productifs en raison des inondations et des glissements de terrain qui s'ensuivent. Néanmoins, à ce jour, les patrons de circulation climatique clé de leurs évolutions et magnitudes différentes sont encore très peu connus. Dans cette thèse, nous montrons que les différentes configurations de précipitations lors de ces événements sont associées au transport de l'humidité provenant de différentes sources d'humidité à grande échelle. Lors des événements 1983-1998 (2016), ils sont respectivement liés à une forte (faible à modérée) humidité provenant du réchauffement du Pacifique (également mais aussi provenant de L'Atlantique via l'Amazonie). La caractéristique de ces transports d'humidité est due à une réponse atmosphérique opposée entre les événements de 1983-1998 (qui sont similaires) et l'événement de 2016 qui présente des patterns de transport d'humidité déphasés. Ces précipitations sont liées à l'humidité provenant de ces sources et la circulation atmosphérique régionale des vents de niveau supérieur (100 à 300 hPa) influe sur la quantité d'humidité qui pénètre dans la région Nord-Centre du versant Pacifique Péruvien. L'interaction de la circulation à grande échelle et régionale et le transport de l'humidité du Pacifique est expliqué par le mode Ep qui est associé à des précipitations dans la région Centre-Nord. La forte dispersion des précipitations dans les régions montagneuses est expliquée par le mode Cp pendant les phénomènes El Niño modérés (extrêmes) et est liée au transport d'humidité de niveau moyen-bas (haut) de l'Amazonie (Pacifique) atteignant les hautes terres. / Four strong El Niño events took place within the last five decades (1972/1973, 1982/1983, 1997/1998 and 2015/2016) recorded as strong in the Niño 3.4 region. They can exhibit significant differences in their evolution associated with a distinct rainfall anomaly evolution along the PPB (Peruvian Pacific Basin), which illustrates the strong nonlinearity of the ENSO teleconnection on the rainfall in this area. These extreme rainfalls have harmful impacts on the population and productive sectors due to floods and landslides which are trigged by them. Yet the key climatic circulation pattern for their different evolution and magnitude are still unknown. Here we show that different rainfall patterns during these events are associated with moisture transport originated from different large-scale moisture sources. For example, in the 1983 -1998 (2016) events appear as related with strong (weak to moderate) moisture coming from the Pacific warming (also coming from Atlantic Ocean through the Amazon basin). Characteristic of these moisture transports is due to an atmospheric response opposite between the 1983-1998 events (that are similar) and 2016 event experiencing out-of-phase moisture transport patterns. Although these rainfalls are linked to the moisture arrival from those sources, the moisture amount entering the PPB can be also influenced by regional atmospheric circulation of upper level winds (100 to 300 hPa) leading to different enhanced moisture transport associated with different rainfall anomalies in the North-Centre PPB. The interplay of large-scale and regional circulation and Pacific moisture transport explains the Ep mode associated with rainfall in the north-Centre PPB. The high dispersion of rainfall in highlands (Cp mode) during the moderate (extremes) El Niño appears as linked to low-middle (high level) moisture transport from the Amazon (Pacific) reaching highlands.

L'oscillation de Madden-Julian (MJO) est la composante dominante de la variabilité intrasaisonnière dans l'atmosphère tropicale, se propageant vers l'est dans la bande équatoriale. Elle se compose d'un centre convectif (phase active) accompagné de la convergence des anomalies du vent zonal de bas niveau et de la divergence de niveau supérieur, et de zones de convection faible (phases supprimées). Trois périodes de l'activité MJO sur l'océan Indien et le continent maritime ont été choisies : 6-14 avril 2009, 23-30 novembre 2011 et 9-28 février 2013. Les simulations avec et sans paramétrisation de la convection ont été réalisées pour un grand domaine avec le modèle atmosphérique Méso-NH. Il a été obtenu que les simulations avec convection paramétrée n'étaient pas capables de reproduire un signal MJO. Pour 2009 et 2011, lorsque le couplage entre la convection et la circulation de grande échelle était fort, les simulations avec convection explicite ont montré une propagation visible de la MJO, ce qui n'a pas été le cas pour 2013. Pour 2011, les processus contribuant à la suppression de la convection ont été étudiés avec une analyse isentropique pour séparer les masses d'air ascendantes ayant une température potentielle équivalente élevée des masses d'air subsidentes ayant une température potentielle équivalente faible. Trois circulations de grande échelle ont été trouvées : une circulation troposphérique, une circulation de percées nuageuses dans la couche de tropopause tropicale, et une circulation de masses d'air à faible température potentielle équivalente dans la basse troposphère. Cette dernière correspond aux intrusions d'air sec de grande échelle des zones subtropicales dans la bande équatoriale, trouvées principalement pendant la phase supprimée de la MJO sur l'océan Indien. / The Madden-Julian Oscillation (MJO) is the dominant component of the intraseasonal variability in the tropical atmosphere, propagating eastward in the equatorial band. It consists of a convective center (active phase) accompanied by the low-level zonal wind anomaly convergence and the upper-level zonal wind anomaly divergence, and zones of weak convection (suppressed phases). Three time periods of the MJO activity over the Indian Ocean and the Maritime Continent were chosen: 6-14 April 2009, 23-30 November 2011, and 9-28 February 2013. The simulations with and without convective parameterizations were performed for a large domain with the atmospheric model Méso-NH. It was obtained that the simulations with parameterized convection were not able to reproduce an MJO signal. For 2009 and 2011 when the coupling between convection and large-scale circulation was strong, the convection-permitting simulations showed a visible MJO propagation, which was not the case for 2013. For the 2011 episode, the processes contributing to the suppression of the convection were studied using an isentropic analysis to separate the ascending air masses with high equivalent potential temperature from the subsiding air masses with low equivalent potential temperature. Three large-scale circulations were found: a tropospheric circulation, an overshoot circulation within the tropical tropopause layer, and a circulation of air masses with low equivalent potential temperature in the lower troposphere. The latter corresponds to the large-scale dry air intrusions from the subtropical zones into the equatorial band, mostly found during the suppressed MJO phase over the Indian Ocean.

Le lac Tonle Sap (TSL) est l'un des plus grands lacs du monde et est connu comme un hot-spot de la biodiversité en Asie du Sud-Est. En raison de la grande diversité ainsi que de la productivité élevée de poissons, le lac contribue à un apport important en protéines pour la population cambodgienne. Cette étude vise principalement à étudier (1) la dynamique spatiale et temporelle de la composition des communautés de poissons, (2) les effets des facteurs environnementaux sur la distribution des espèces et (3) l'effet de l'abolition des lots de pêche sur la biomasse, les communautés de poissions et le structure des poissons. En utilisant des méthodes statistiques multivariées sur les données de poissons et de variables environnementales, je suis en mesure de mettre en évidence les principales conclusions suivantes: * Les communautés de poissons du TSL étaient composées de deux assemblages de poissons: l'assemblage du nord, principalement caractérisé par des poissons noirs, et l'assemblage du sud, principalement lié aux poissons blancs, gris et estuariens. Les assemblages de poissons de la période 1994-1995 étaient représentés par l'abondance de tous les groupes fonctionnels, c'est-à-dire les poissons noirs, blancs et gris, et pour la période de 1996 à 1999, les assemblages étaient liés aux poissons blancs et gris. * Les distributions des abondances espèces de poissons n'étaient pas homogènes dans le TSL. De plus, les aires de distributions des espèces étaient différentes et étaient régies par des combinaisons distinctes de caractéristiques de l'habitat et de facteurs climatiques. * H. lobatus et H. siamensis peuvent coexister ensemble, mais la synchronisation et la migration de H. lobatus conduisent toujours à celles de H. siamensis. Ces résultats suggèrent que la population de H. lobatus est plus sensible aux variations d'impulsion de flux que celles de H. siamensis. Ceci indique que les variations des impulsions d'écoulement sont les principaux déterminants responsables de la dynamique temporelle de chaque espèce.[...] / Tonle Sap Lake (TSL) is one of the world's largest lakes and is a biodiversity hotspot in Southeast Asia. It supports high fish productivity which sustains protein supply for millions of people in the region. This study aims to investigate (1) spatial and temporal dynamics of fish community composition, (2) the effects of environmental factors on fish distribution and (3) effects of fishing lot abolishment on fish biomass, community and structure in TSL. By using multivariate statistical methods on fish and environmental data, the thesis highlights that: * There were two fish assemblages in TSL: the northern assemblage, mostly characterised by black fishes, and the southern assemblage, mainly linked to white, grey and estuarine fishes. Fish assemblages from earlier years (1994 and 1995) were represented by the abundances of all functional groups, i.e. black, white and grey fishes, and from 1996 to 1999, the assemblages were linked only to white and grey fishes. * Fish species distributions were not homogeneous within TSL. In addition, species distribution areas were different and were governed by distinct combinations of the local habitat characteristics and regional climatic factors. * H. lobatus and H. siamensis can co-occur together, but synchronisation and migration of H. lobatus always lead those of H. siamensis. These results suggest that the population of H. lobatus is more responsive to flow pulse variations than those of H. siamensis.[...]

No description available.

Avec l'émergence de l'informatique dans les nuages, une nouvelle approche consiste à externaliser des tâches de calcul, de façon à réduire les coûts d'hébergement et à augmenter la flexibilité des systèmes. L'infrastructure actuelle des services permettant cette externalisation repose sur l'utilisation de centres de traitement de données centralisés, qui sont dédiés à l'approvisionnement de ressources de calcul. Dans cette thèse, nous étudions la possibilité de fournir de tels services en utilisant une infrastructure pair-à-pair, c'est-à-dire une infrastructure totalement décentralisée pouvant être déployée sur une fédération de noeuds de calcul hétérogénes et de provenances diverses. Nous nous focalisons sur le problème de l'allocation des noeuds et présentons Salute, un service d'allocation de noeuds, qui organise les noeuds en réseaux virtuels non-structurés et repose sur des mécanismes de prédiction de disponibilité pour assurer, avec une grande probabilité, que les requêtes d'allocation sont satisfaites dans le temps, malgré le dynamisme de l'environnement hôte. Pour ce faire, le service Salute repose sur la collaboration de plusieurs protocoles pair-à-pair appartenant à la catégorie des protocoles épidémiques. Afin de valider nos propositions, nous évaluons Salute en utilisant des traces provenant d'un échantillonnage de plusieurs systèmes pair-à-pair de référence. / With the emergence of Cloud computing, a new trend is to externalize computing tasks in order to decrease costs and increase flexibility. Current Cloud infrastructures rely on the usage of large-scale centralized data centers, for computing resources provisioning. In this thesis, we study the possibility to provide a peer-to-peer based Cloud infrastructure, which is totally decentralized and can be deployed on any computing nodes federation. We focus on the nodes allocation problem and present Salute, a nodes allocation service that organizes nodes in unstructured overlay networks and relies on mechanisms to predict node availability in order to ensure, with high probability, that allocation requests will be satisfied over time, and this despite churn. Salute's implementation relies on the collaboration of several peer-to-peer protocols belonging to the category of epidemic protocols. To convey our claims, we evaluate Salute using real traces.