Devido ao processo de urbanização sem planejamento adequado, as cidades costeiras sofrem com as inundações e erosão das praias com a elevação do nível do mar e ação de ondas. Entre as estruturas mais utilizadas para proteção de praias destacam-se os quebra-mares segmentados. Para estudar e prever as respostas praiais à presença destas estruturas, podem ser empregados modelos numéricos, capazes de calcular os processos complexos deste ambiente. Este trabalho teve como objetivo o estudo da resposta hidrodinâmica e sedimentar de uma praia com a implantação de quebra-mares segmentados. O modelo hidrodinâmico utilizado é o MIKE21FM acoplado ao modelo de ondas SW e sedimentos ST e SM. Primeiramente, foram impostas condições hidrodinâmicas simplificadas e diferentes arranjos de quebra-mares segmentados. As áreas erodidas e acrescidas após a implantação da obra foram comparadas e priorizou-se minimizar a área erodida. Em seguida, foram considerados os cenários de condições hidrodinâmicas variadas com uma das estruturas. Os resultados são apresentados em termos de altura significativa e potência de onda para 18 casos de condições hidrodinâmicas constantes, as quais foram feitas variar em nível e características da onda. Para a condição hidrodinâmica mais crítica, obteve-se uma redução de 0,4m da altura significativa média e 2kW/m da potência média da onda. Através desses resultados é possível aquilatar a proteção da região da praia da ação das ondas após a implantação da estrutura. Além disso, foram feitas duas simulações com 3 a 4 meses de evolução do fundo em condições hidrodinâmicas reais, com o intuito de apresentar qualitativamente as áreas de erosão e acreção do fundo da praia. Resultados mostraram que a deriva litorânea local tem direção sudeste, para a Ponta da Praia. Foi observada sedimentação no tardoz dos segmentos e na região de sombra e erosão nas áreas entre segmentos e a frente destes. Com isso, este trabalho colabora com o gerenciamento costeiro na região de Santos, apresentando o início de uma pesquisa para uma possível solução para os problemas de inundação e erosão da região da Ponta da Praia. / Erosion and flooding are common problems faced by coastal cities as a consequence of urbanization process without previous planning, with the sea level rise and the wave action. In order to mitigate these problems, shoreline protection structures are constructed along the coastlines, such as segmented breakwaters. Numerical models are tools frequently used in these studies because of its capacity of calculating the complex processes on a coastal environment. This study aim is to present a study based on the hydrodynamic and sediment transport response of a curved beach to the presence of segmented emerged breakwaters. The numerical model used is the MIKE21FM coupled with wave spectral and sediment transport modules. Firstly, constant boundary conditions were applied for the period of one year in order to understand the response of the beach to five different designs of coastal breakwaters. One of these structures were chosen to continue the study based on the area and length of the beach eroded and accreted. Secondly, eighteen different scenarios of wave and water level were applied to the selected structure and the protection of wave action were compared and analyzed. To the most critical hydrodynamic condition, the structure reduced 0,4 m of the mean significant wave height and 2 kW/m of the mean wave power. Finally, two real hydrodynamic conditions were applied and the sediment transport was qualitatively analyzed. Results show that the littoral drift in the studied area is directed to the southeast end of the beach, straight to the Ponta da Praia, without the spur breakwater, the sediment would end inside the navigation channel. Besides that, it shows the sedimentation and erosion patterns at the beach generated by the presence of the breakwater. The sediment is accumulated at the shadow area and at the lee, whereas the erosion is located at the front of the breakwaters and between the segments, as expected. Therefore, this study collaborates with the management of Santos, by presenting herewith a possible solution for the inundation and erosion problems of this area.

O crescimento da produção de petróleo offshore traz consigo a demanda por tubulações para escoamento de hidrocarbonetos capazes de conciliar elevada resistência mecânica, elevada resistência à corrosão e baixo peso estrutural. Para tal, após o projeto mecânico, faz-se necessária a qualificação dessas estruturas por intermédio de ensaios mecânicos. Um dos métodos de teste em risers rígidos que alia baixos custos em equipamentos à rapidez de execução é o teste de fadiga por ressonância. Usualmente, este tipo de teste é controlado através do nível de deformação lido por extensômetros instalados na região central da amostra. Contudo, o tempo de instalação e a fragilidade dos sensores acabam por aumentar significativamente o tempo total de teste. Portanto, o presente estudo propõe que o controle do teste seja feito por intermédio dos níveis de aceleração no centro do tubo, uma vez que acelerômetros são mais robustos e mais fáceis de instalar. Assim, o trabalho traz uma correlação numérica e experimental entre os níveis de deformação e aceleração no centro de tubos submetidos a testes de fadiga por ressonância. Foram realizados testes nas amostras sem água, com água, e com pressão interna de 200 Bar. Foi constatado que a relação entre a aceleração e a deformação no centro das amostras para frequências próximas à frequência de ressonância é linear. Além disso, as diferenças observadas entre os resultados experimentais e os numéricos variaram de 0,13% a 14,73%. Em conclusão, o modelo numérico apresentou resultados consistentes em grande parte dos casos estudados, o que evidencia a viabilidade de implementação do sistema de controle por aceleração. Para tal, foram apresentadas sugestões de melhoria para aplicação em testes futuros, a fim de otimizar a uniformidade dos resultados e completar a validação do modelo numérico. / The development of offshore oil production implies the demand of pipelines for the hydrocarbon flow, which are capable of combining high mechanical and corrosion resistance with low structural weight. For this, after the mechanical design, it is required to qualify these structures by means of mechanical tests. One method of test for rigid risers that combines low equipment costs with quick execution is the resonance fatigue test. Usually, this type of test is controlled by strain levels obtained by strain gauges installed at the pipe center. However, the installation lead time and the sensors fragility increase significantly the overall test time. Therefore, the present study proposes that the test controlling system could be performed by means of the acceleration levels at the middle of the sample, since accelerometers are more robust and the installation is simpler. Thus, this work brings a numerical and experimental relationship between strain and acceleration levels at the center of pipes submitted to resonance fatigue test. Tests were performed on the samples without water, with water, and with internal pressure of 200 Bar. It was found that the relationship between strain and acceleration at the center of the samples is linear for frequencies near the resonant frequency. In addition, the differences between experimental and numerical results ranged from 0.13% to 14.73%. In conclusion, the numerical model presented consistent results in most of the studied cases, which evidences the feasibility of implementing the acceleration controlling system. For this purpose, improvement suggestions for application in future tests were presented in order to optimize the uniformity of the results and to complete the numerical model validation.

Dados de altura da superfície do mar medidos a partir de satélites altimétricos e saídas de dois modelos de circulação geral dos oceanos (OGCM for the Earth Simulator - OFES e Community Earth System Model - CESM) foram utilizados com o objetivo de verificar se modelos numéricos climáticos reproduzem ondas de Rossby de forma similar à observada em dados altimétricos no Atlântico Sul. Os modelos diferem quanto à forçante, o modelo OFES é forçado com dados do NECP-NCAR enquanto que o modelo CESM é acoplado com modelo atmosférico, de gelo e terrestre. Uma vez que essas ondas dependem da estrutura interna de densidade e da forçante do vento, podemos fazer inferências sobre a adequação desses dois fatores nos modelos à realidade observada pelo altímetro. Uma série de filtros de resposta impulsiva finita 2D (FIR-2D) foi aplicada aos dados de anomalia da altura da superfície do mar com o propósito de detectar as ondas de Rossby e as componentes de onda encontradas nos dados dos modelos foram comparadas com as do altímetro. Os dois modelos são capazes de reproduzir ondas de Rossby e apresentam melhores resultados em baixas latitudes. Porém o modelo OFES apresentou mais dificuldades para reproduzir os parâmetros de onda encontrados no altímetro apresentando diferenças médias de até 68% para a amplitude, 34% para o comprimento de onda e 38% para a velocidade de fase. Em contrapartida as maiores diferenças médias entre os parâmetros de onda calculados a partir dos dados altimétricos e do modelo CESM foram de 32% para a amplitude, 20% para o comprimento de onda e 20% para a velocidade de fase. Além disso, o modelo CESM foi capaz de reproduzir o sinal sazonal com uma correlação média de aproximadamente 0,7 com o sinal sazonal encontrado pelo altímetro em toda a bacia do oceano Atlântico Sul enquanto que o sinal sazonal do modelo OFES apresentou uma correlação média de 0,4 com o sinal encontrado pelo satélite. Estes resultados mostram que os dois modelos reproduzem o fenômeno satisfatoriamente, sendo o CESM melhor que o OFES. A diferença dos resultados deve estar ligada aos aspectos supracitados, especificamente à forçante atmosférica e estrutura de densidade na coluna d\'água nas latitudes ao sul de 20ºS. / Sea surface height data measured from altimetry satellites and outputs of two ocean general circulation models (OGCM for the Earth Simulator - OFES and Community Earth System Model - CESM) were used to determine whether climate numerical models reproduce the Rossby waves in a manner similar to those observed in altimetry record in the South Atlantic. The models differs in the forcing, OFES is forced with NCEP-NCAR data while CESM is coupled with atmospheric model, ice and land model. Because these waves depend on the internal density structure and wind forcing, we can make inferences about the suitability of these two factors in the models in comparison with the altimetry data used as a reference. A series of finite impulse response band-pass filters (FIR-2D) was applied to isolate the westward propagating signals corresponding to Rossby waves in the altimeter. Both models are able to reproduce Rossby waves and show better results in low latitudes. However, the OFES model presented more difficulty to reproduce the wave parameters found in altimeter with differences of up to 68% for amplitude, 34% for the wavelength and 38% for phase velocity. By contrast the greatest differences between the wave parameters computed from the altimeter data and the CESM model were 32% for amplitude, 20% for the wavelength and 20% for phase velocity. Furthermore, the CESM model was capable of reproducing the seasonal signal correlation with an average of approximately 0.7 with altimeter\'s seasonal signal found throughout the South Atlantic basin, while the OFES\'s seasonal signal showed an average correlation 0.4 with the signal found by the satellite. These results indicate that both models can reproduce the phenomenon satisfactorily, the CESM better than OFES. The difference between the results should be related with the aspects cited above, specifically with the atmospheric forcing and the density structure in the water column in latitudes southern then 20ºS.

L'objectif principal de notre projet est de créer un modèle numérique complet du système pelvien de la femme enceinte prenant en compte l'intégralité des structures anatomiques de suspensions du pelvis, leurs propriétés biomécaniques et permettant la simulation d'accouchement eutocique et dystocique.Nous travaillons avec un modèle tridimensionel du pelvis de la femme enceinte créé à partir d'IRM réalisées chez une parturiente volontaire à différents âges gestationnels de la grossesse (16, 32 et 38SA) et en post-partum (2mois, 1an). Chaque organes, muscles et ligaments du pelvis ont été reconstruits numériquement permettant de générer un modèle géométrique complet représentatif du système pelvien. Nous avons effectué une analyse géométrique de ces structures, notamment musculaires et ligamentaires. Par la suite nous avons effectué des simulations numériques par méthode des Eléments Finis (EF), en intégrant notamment les propriétés biomécaniques des tissus et des trajectoires de tête imposées. Ce système permet d'analyser les déformations des structures périnéales lors d'un accouchement normal puis dystocique.L'étude approfondie au cours de la grossesse des ligaments utérosacrés et du muscle releveur de l'anus, met en évidence une modification géométrique de ces structures, jouant un rôle dans la statique pelvienne et ce dès le 2ème trimestre de grossesse. Cette reconstruction numérique en 3D a permis de réaliser un modèle d'enseignement (mannequin simple) et d'entraînement au geste de délivrance artificielle et de révision utérine. Cet outil a été testés auprès de gynécologue-obstétriciens, sages-femmes et internes afin d'élaborer un modèle final dont l'intérêt pédagogique a été validé. Par la suite, nous avons élaboré un modèle de simulation d'accouchement, paramétré, permettant de faire varier différents critères maternels et foetaux comme l'orientation, la taille et la flexion de la tête foetale. Dans un premier temps la simulation d'accouchement physiologique permet l'analyse de l'impact de l'accouchement sur ces mêmes structures pelviennes. Nous avons dans un second temps fait varier différents critères permettant d'aboutir à la dystocie. Nous travaillons également sur le développement de capteurs de pressions et mouvements, intégrés à des cuillères de forceps. Cet outil permet d'effectuer une analyse numérique du geste d'extraction par forceps.Cet outil paramétré de simulation numérique d'accouchement va permettre d'étudier les différents impacts de l'accouchement sur l'équilibre fragile des structures pelviennes. Ainsi il serait possible d'analyser la survenue de complications obstétricales maternelles et foetales. Ainsi il serait possible d'expliquer les lésions périnéales à court, moyen et long terme lié à l'accouchement. Cet outil permettrait également d'enseigner la mécanique et la gestuelle obstétricale parfois complexe. / We aim at developing a complete 3D numerical model of a parturient pelvic system representing all the anatomical structures of the pelvis such as ligament, muscle and organs. Then we generate a parametric FE model that allows simulating normal and dystocic vaginal delivery.We have developed a parturient pelvic numerical model at different gestational ages, 16, 32 and 38 weeks of gestation, (WG) and in postpartum (2months and 1year) from MRI. The different organs, muscles and ligaments of the pelvic system were segmented in order to generate a complete anatomical 3D model. Starting from this numerical model we studied the changes the muscles and ligaments undergo during pregnancy. Then we performed a Finite Element (FE) model that allows simulation and analysis of the deformations of pelvic anatomical structures under the stress induce by normal and dystocic vaginal delivery. In particular, we investigated the influence of the head size, terms and cephalic orientation and flexion. We particularly studied the structures that play an important role in the stability of the pelvic system.ResultsThe analysis during pregnancy of the US ligaments and levator ani muscle (LAM) reveals some geometrical modification, even then at the beginning of the second pregnancy trimester. This 3D anatomical model help to develop a teaching model for manual removing of the placenta, that could be integrated in a simple physic mannequin. The proof of pedagogical interest of this tool was made by different series of tests, underwent by gynaecolog-obstetrician and midwives. Then we worked with FE simulation of the vaginal delivery. The model developed is parametric. Than mean we can then change different maternal and fetal criteria such as gestational age, fetal head size, orientation and flexion. First place we performed normal vaginal delivery to study the impact of the fetal head descent in the pelvic system, and his stress impact on the different anatomical structures. Then we introduce dystocic element. We can evaluate and localize the strain levels and the most injured areas. Posterior cephalic presentation presents higher injury risk than the anterior one. Maternal geometry at different terms brings equivalent results contrary to the fetal head sizes that have an influence on the strain level and the potential damage induced. We developed pressure and trajectories sensors integrated in a forceps. We can then record an ex-vivo forceps extraction and then integrate all the information in the FE model.ConclusionThis multi-parametric investigation allows us to have a customizable and predictive tool evaluating the potential damages on the pelvis during vaginal delivery. We could then explain, understand and maybe predict some maternal and fetal complications that could happen during vaginal delivery. We can in particular try to explain the perinea injuries during, after and long time after vaginal delivery. This tool can be used to teach the complexity of obstetric.

In this thesis, we produce a rigorous and quantitative analysis of the errors introduced by finite difference schemes into strong constraint 4D-Variational (4D-Var) data assimilation. Strong constraint 4D-Var data assimilation is a method that solves a particular kind of inverse problem; given a set of observations and a numerical model for a physical system together with a priori information on the initial condition, estimate an improved initial condition for the numerical model, known as the analysis vector. This method has many forms of error affecting the accuracy of the analysis vector, and is derived under the assumption that the numerical model is perfect, when in reality this is not true. Therefore it is important to assess whether this assumption is realistic and if not, how the method should be modified to account for model error. Here we analyse how the errors introduced by finite difference schemes used as the numerical model, affect the accuracy of the analysis vector. Initially the 1D linear advection equation is considered as our physical system. All forms of error, other than those introduced by finite difference schemes, are initially removed. The error introduced by `representative schemes' is considered in terms of numerical dissipation and numerical dispersion. A spectral approach is successfully implemented to analyse the impact on the analysis vector, examining the effects on unresolvable wavenumber components and the l2-norm of the error. Subsequently, a similar also successful analysis is conducted when observation errors are re-introduced to the problem. We then explore how the results can be extended to weak constraint 4D-Var. The 2D linear advection equation is then considered as our physical system, demonstrating how the results from the 1D problem extend to 2D. The linearised shallow water equations extend the problem further, highlighting the difficulties associated with analysing a coupled system of PDEs.

Fire is a significant threat to the structural integrity of buildings. Depending on the architecture of the structure and the intensity and duration of the fire event, structural members may lose strength and stiffness eventually leading to collapse whether by flexural buckling or crushing. The focus of this research is on the behavior and fire performance of reinforced-concrete columns under fire conditions. In order to effectively study column performance with differing loading, aggregate and dimensional characteristics under varying time-temperature curves and fire exposures, a numerical model was constructed in Microsoft Excel. The spreadsheet model allowed for complete transparency of the calculations and provided a means to visualize the data in flexible ways. ANSYS and several published column furnace tests were used to benchmark the heat transfer and structural analysis portions of the model. One, three and four-sided fire exposures along with the ASTM E119 fire curve and a natural fire curve were used to study latent heating effects, increasing and decreasing eccentricities, moment magnification, and failure modes. Assessments of column structural capacity were performed in accordance with the provisions of ACI 318. The completed model served as an effective tool for the thesis and is available to help aid students and engineers investigate the design of reinforced concrete columns under fire conditions through integration the heat transfer analyses and the structural evaluations.

Devido ao processo de urbanização sem planejamento adequado, as cidades costeiras sofrem com as inundações e erosão das praias com a elevação do nível do mar e ação de ondas. Entre as estruturas mais utilizadas para proteção de praias destacam-se os quebra-mares segmentados. Para estudar e prever as respostas praiais à presença destas estruturas, podem ser empregados modelos numéricos, capazes de calcular os processos complexos deste ambiente. Este trabalho teve como objetivo o estudo da resposta hidrodinâmica e sedimentar de uma praia com a implantação de quebra-mares segmentados. O modelo hidrodinâmico utilizado é o MIKE21FM acoplado ao modelo de ondas SW e sedimentos ST e SM. Primeiramente, foram impostas condições hidrodinâmicas simplificadas e diferentes arranjos de quebra-mares segmentados. As áreas erodidas e acrescidas após a implantação da obra foram comparadas e priorizou-se minimizar a área erodida. Em seguida, foram considerados os cenários de condições hidrodinâmicas variadas com uma das estruturas. Os resultados são apresentados em termos de altura significativa e potência de onda para 18 casos de condições hidrodinâmicas constantes, as quais foram feitas variar em nível e características da onda. Para a condição hidrodinâmica mais crítica, obteve-se uma redução de 0,4m da altura significativa média e 2kW/m da potência média da onda. Através desses resultados é possível aquilatar a proteção da região da praia da ação das ondas após a implantação da estrutura. Além disso, foram feitas duas simulações com 3 a 4 meses de evolução do fundo em condições hidrodinâmicas reais, com o intuito de apresentar qualitativamente as áreas de erosão e acreção do fundo da praia. Resultados mostraram que a deriva litorânea local tem direção sudeste, para a Ponta da Praia. Foi observada sedimentação no tardoz dos segmentos e na região de sombra e erosão nas áreas entre segmentos e a frente destes. Com isso, este trabalho colabora com o gerenciamento costeiro na região de Santos, apresentando o início de uma pesquisa para uma possível solução para os problemas de inundação e erosão da região da Ponta da Praia. / Erosion and flooding are common problems faced by coastal cities as a consequence of urbanization process without previous planning, with the sea level rise and the wave action. In order to mitigate these problems, shoreline protection structures are constructed along the coastlines, such as segmented breakwaters. Numerical models are tools frequently used in these studies because of its capacity of calculating the complex processes on a coastal environment. This study aim is to present a study based on the hydrodynamic and sediment transport response of a curved beach to the presence of segmented emerged breakwaters. The numerical model used is the MIKE21FM coupled with wave spectral and sediment transport modules. Firstly, constant boundary conditions were applied for the period of one year in order to understand the response of the beach to five different designs of coastal breakwaters. One of these structures were chosen to continue the study based on the area and length of the beach eroded and accreted. Secondly, eighteen different scenarios of wave and water level were applied to the selected structure and the protection of wave action were compared and analyzed. To the most critical hydrodynamic condition, the structure reduced 0,4 m of the mean significant wave height and 2 kW/m of the mean wave power. Finally, two real hydrodynamic conditions were applied and the sediment transport was qualitatively analyzed. Results show that the littoral drift in the studied area is directed to the southeast end of the beach, straight to the Ponta da Praia, without the spur breakwater, the sediment would end inside the navigation channel. Besides that, it shows the sedimentation and erosion patterns at the beach generated by the presence of the breakwater. The sediment is accumulated at the shadow area and at the lee, whereas the erosion is located at the front of the breakwaters and between the segments, as expected. Therefore, this study collaborates with the management of Santos, by presenting herewith a possible solution for the inundation and erosion problems of this area.

This Master´s Thesis studies the applicability of existing numerical model to predict the conditions in the drifts at the mine. Damage mapping, covering the existing production levels, has been carried out to study the condition of the surface support, i.e. shotcrete, mesh and rock bolts to quantify the amount of deformation visible in the tunnels. Irregularities in the floor levels were also mapped. The rock support applied varies between different areas, from only a layer of shotcrete to areas where shotcrete, mesh and cable bolts are applied sequentially. The large amount of support in some regions is required because of high in situ rock stresses in Kemi Mine. The geological conditions are challenging, with large local variation making efficient supporting and damage prediction difficult. Access drifts at the mine have varying service times, on average 6 years, which is a long time in a difficult environment. The results from this work are a mine specific damage classification, used in the mapping to capture the range of damages seen on site. A reoccurrence of same areas showing damages on several production levels is noted. Digitized versions of the damage maps were made and these have been compared to simulation results. The comparisons were made to plots of deviatoric strain, deviatoric stress and total displacement. Based on the work done during this project it can be concluded that the studied parameters have varying levels of agreement with the drifts, both when comparing the levels with each other and the results in the same level between spring and autumn. The best agreement is found with the deviatoric strain increment and the yielded elements. Rock mechanics and geological reasons for the variations seen between the mapping results and the simulations results can be further studied in the future, as can the alternatives for increasing the accuracy through changes in the numerical model or model type used for comparison.

Dados de altura da superfície do mar medidos a partir de satélites altimétricos e saídas de dois modelos de circulação geral dos oceanos (OGCM for the Earth Simulator - OFES e Community Earth System Model - CESM) foram utilizados com o objetivo de verificar se modelos numéricos climáticos reproduzem ondas de Rossby de forma similar à observada em dados altimétricos no Atlântico Sul. Os modelos diferem quanto à forçante, o modelo OFES é forçado com dados do NECP-NCAR enquanto que o modelo CESM é acoplado com modelo atmosférico, de gelo e terrestre. Uma vez que essas ondas dependem da estrutura interna de densidade e da forçante do vento, podemos fazer inferências sobre a adequação desses dois fatores nos modelos à realidade observada pelo altímetro. Uma série de filtros de resposta impulsiva finita 2D (FIR-2D) foi aplicada aos dados de anomalia da altura da superfície do mar com o propósito de detectar as ondas de Rossby e as componentes de onda encontradas nos dados dos modelos foram comparadas com as do altímetro. Os dois modelos são capazes de reproduzir ondas de Rossby e apresentam melhores resultados em baixas latitudes. Porém o modelo OFES apresentou mais dificuldades para reproduzir os parâmetros de onda encontrados no altímetro apresentando diferenças médias de até 68% para a amplitude, 34% para o comprimento de onda e 38% para a velocidade de fase. Em contrapartida as maiores diferenças médias entre os parâmetros de onda calculados a partir dos dados altimétricos e do modelo CESM foram de 32% para a amplitude, 20% para o comprimento de onda e 20% para a velocidade de fase. Além disso, o modelo CESM foi capaz de reproduzir o sinal sazonal com uma correlação média de aproximadamente 0,7 com o sinal sazonal encontrado pelo altímetro em toda a bacia do oceano Atlântico Sul enquanto que o sinal sazonal do modelo OFES apresentou uma correlação média de 0,4 com o sinal encontrado pelo satélite. Estes resultados mostram que os dois modelos reproduzem o fenômeno satisfatoriamente, sendo o CESM melhor que o OFES. A diferença dos resultados deve estar ligada aos aspectos supracitados, especificamente à forçante atmosférica e estrutura de densidade na coluna d\'água nas latitudes ao sul de 20ºS. / Sea surface height data measured from altimetry satellites and outputs of two ocean general circulation models (OGCM for the Earth Simulator - OFES and Community Earth System Model - CESM) were used to determine whether climate numerical models reproduce the Rossby waves in a manner similar to those observed in altimetry record in the South Atlantic. The models differs in the forcing, OFES is forced with NCEP-NCAR data while CESM is coupled with atmospheric model, ice and land model. Because these waves depend on the internal density structure and wind forcing, we can make inferences about the suitability of these two factors in the models in comparison with the altimetry data used as a reference. A series of finite impulse response band-pass filters (FIR-2D) was applied to isolate the westward propagating signals corresponding to Rossby waves in the altimeter. Both models are able to reproduce Rossby waves and show better results in low latitudes. However, the OFES model presented more difficulty to reproduce the wave parameters found in altimeter with differences of up to 68% for amplitude, 34% for the wavelength and 38% for phase velocity. By contrast the greatest differences between the wave parameters computed from the altimeter data and the CESM model were 32% for amplitude, 20% for the wavelength and 20% for phase velocity. Furthermore, the CESM model was capable of reproducing the seasonal signal correlation with an average of approximately 0.7 with altimeter\'s seasonal signal found throughout the South Atlantic basin, while the OFES\'s seasonal signal showed an average correlation 0.4 with the signal found by the satellite. These results indicate that both models can reproduce the phenomenon satisfactorily, the CESM better than OFES. The difference between the results should be related with the aspects cited above, specifically with the atmospheric forcing and the density structure in the water column in latitudes southern then 20ºS.

O reparo em dutos defeituosos é previsto por normas e literatura, porém existe grande espaço para definição dos parâmetros ideais para aplicação prática e avaliação do desempenho destes reparos em fadiga. O presente trabalho tem por objetivo a definição dos parâmetros ideais para aplicação de reparos eficientes do tipo dupla calha sem solda circunferencial e com carregamento externo em escala real em dutos com defeitos de mossa com cava. O reparo é baseado na instalação de duas calhas concêntricas ao duto a ser reparado, preenchendo-se o defeito com resina e aplicando carga de compressão através de atuadores hidráulicos com correntes a fim de induzir tensões compressivas na região do defeito e, por fim, as calhas são unidas longitudinalmente pelo processo de soldagem. Para este trabalho, foi realizada uma modelagem numérica utilizando o método de elementos finitos de modo a simular a reprodução do defeito, ensaios de fadiga sem reparo, parâmetros para aplicação do reparo, ensaios de fadiga com reparo e, por fim, ensaio hidrostático. O procedimento proposto foi a reprodução do defeito, primeiramente o duto foi deformado com uma esfera metálica, de modo a formar uma mossa de profundidade igual a 6 % do diâmetro externo e, em seguida, a cava foi usinada com a utilização de um disco abrasivo. Após a produção do defeito, três amostras foram ensaiadas em fadiga, de modo a produzir uma amostra de referência sem reparo a ser comparada com os corpos de prova com reparo. O reparo dupla calha foi aplicado em duas amostras com defeito, sendo essas então submetidas a ensaios de fadiga. Esses ensaios indicaram que ambas as amostras resistiram a um número de ciclos correspondentes a 100 anos de vida em serviço. Por fim, as amostras reparadas após o ensaio de fadiga foram submetidas a ensaios hidrostáticos até a ruptura, a fim de avaliar a vida remanescente. Ambos os dutos romperam fora da região reparada, com uma pressão 2,1 %, menor que a pressão calculada analiticamente. Como principal resultado, foram obtidos parâmetros ideais para aplicação do reparo dupla calha em dutos com defeito tipo mossa com cava. Os resultados experimentais mostram convergência em relação aos dados do modelo numérico, provando que a modelagem por elementos finitos pode ser uma ferramenta útil no desenvolvimento deste tipo de trabalho. / Defective duct repairs are envisaged by norms and by the literature, yet there is enough space left to define ideal parameters for practical application and performance evaluation of these fatigue repairs. This work’s objective is to define the ideal parameters for the application of efficient repairs in double spouts without circumferential welding and real scale external loading in ducts with dent and gouge defects. The repair is based on the installation of two concentric spouts in the duct, filling the defect with resin and applying compression loading through hydraulic actuators with currents to induce compressive stresses in the defective area. Afterwards, the spouts are united lengthwise through the process of welding. To develop this work, we conducted a numerical modelling by using the finite element method to simulate the reproduction of the defect, as well as fatigue tests with no repairs, parameters for the application of repairs, fatigue tests with repairs, and finally a hydrostatic test. The proposed procedure was the reproduction of the defect. At first the duct was deformed with a metallic sphere, to form a dent whose depth was equal to 6 % of the external diameter and, subsequently, the gouge was machined with the help of an abrasive disk. Following the production of a defect, three samples were tested in fatigue to produce a reference sample without repair which would be compared to the samples with repairs. The double spout repair was applied in two samples with defects, which were then subjected to fatigue tests. These tests showed that both samples resisted to a number of cycles corresponding to a 100-year use lifespan. At last, the samples repaired after fatigue tests were subjected to hydrostatic tests until rupture, in order to evaluate their remaining lifespan. Both ducts ruptured outside of the repaired region, with a stress rate of 2.1 %, lower than the stress calculated analytically. The main result of this study was obtaining ideal parameters for the application of double spout repair in ducts with dents and gouge defects. The experimental results show convergence in relation to the numerical model data, proving that finite element modelling can be a useful tool when conducting this kind of work.