No description available.

In this dissertation, we discuss the problem of robust linear subspace estimation using low-rank optimization and propose three formulations of it. We demonstrate how these formulations can be used to solve fundamental computer vision problems, and provide superior performance in terms of accuracy and running time. Consider a set of observations extracted from images (such as pixel gray values, local features, trajectories . . . etc). If the assumption that these observations are drawn from a liner subspace (or can be linearly approximated) is valid, then the goal is to represent each observation as a linear combination of a compact basis, while maintaining a minimal reconstruction error. One of the earliest, yet most popular, approaches to achieve that is Principal Component Analysis (PCA). However, PCA can only handle Gaussian noise, and thus suffers when the observations are contaminated with gross and sparse outliers. To this end, in this dissertation, we focus on estimating the subspace robustly using low-rank optimization, where the sparse outliers are detected and separated through the `1 norm. The robust estimation has a two-fold advantage: First, the obtained basis better represents the actual subspace because it does not include contributions from the outliers. Second, the detected outliers are often of a specific interest in many applications, as we will show throughout this thesis. We demonstrate four different formulations and applications for low-rank optimization. First, we consider the problem of reconstructing an underwater sequence by removing the iii turbulence caused by the water waves. The main drawback of most previous attempts to tackle this problem is that they heavily depend on modelling the waves, which in fact is ill-posed since the actual behavior of the waves along with the imaging process are complicated and include several noise components; therefore, their results are not satisfactory. In contrast, we propose a novel approach which outperforms the state-of-the-art. The intuition behind our method is that in a sequence where the water is static, the frames would be linearly correlated. Therefore, in the presence of water waves, we may consider the frames as noisy observations drawn from a the subspace of linearly correlated frames. However, the noise introduced by the water waves is not sparse, and thus cannot directly be detected using low-rank optimization. Therefore, we propose a data-driven two-stage approach, where the first stage “sparsifies” the noise, and the second stage detects it. The first stage leverages the temporal mean of the sequence to overcome the structured turbulence of the waves through an iterative registration algorithm. The result of the first stage is a high quality mean and a better structured sequence; however, the sequence still contains unstructured sparse noise. Thus, we employ a second stage at which we extract the sparse errors from the sequence through rank minimization. Our method converges faster, and drastically outperforms state of the art on all testing sequences. Secondly, we consider a closely related situation where an independently moving object is also present in the turbulent video. More precisely, we consider video sequences acquired in a desert battlefields, where atmospheric turbulence is typically present, in addition to independently moving targets. Typical approaches for turbulence mitigation follow averaging or de-warping techniques. Although these methods can reduce the turbulence, they distort the independently moving objects which can often be of great interest. Therefore, we address the iv problem of simultaneous turbulence mitigation and moving object detection. We propose a novel three-term low-rank matrix decomposition approach in which we decompose the turbulence sequence into three components: the background, the turbulence, and the object. We simplify this extremely difficult problem into a minimization of nuclear norm, Frobenius norm, and `1 norm. Our method is based on two observations: First, the turbulence causes dense and Gaussian noise, and therefore can be captured by Frobenius norm, while the moving objects are sparse and thus can be captured by `1 norm. Second, since the object’s motion is linear and intrinsically different than the Gaussian-like turbulence, a Gaussian-based turbulence model can be employed to enforce an additional constraint on the search space of the minimization. We demonstrate the robustness of our approach on challenging sequences which are significantly distorted with atmospheric turbulence and include extremely tiny moving objects. In addition to robustly detecting the subspace of the frames of a sequence, we consider using trajectories as observations in the low-rank optimization framework. In particular, in videos acquired by moving cameras, we track all the pixels in the video and use that to estimate the camera motion subspace. This is particularly useful in activity recognition, which typically requires standard preprocessing steps such as motion compensation, moving object detection, and object tracking. The errors from the motion compensation step propagate to the object detection stage, resulting in miss-detections, which further complicates the tracking stage, resulting in cluttered and incorrect tracks. In contrast, we propose a novel approach which does not follow the standard steps, and accordingly avoids the aforementioned diffi- culties. Our approach is based on Lagrangian particle trajectories which are a set of dense trajectories obtained by advecting optical flow over time, thus capturing the ensemble motions v of a scene. This is done in frames of unaligned video, and no object detection is required. In order to handle the moving camera, we decompose the trajectories into their camera-induced and object-induced components. Having obtained the relevant object motion trajectories, we compute a compact set of chaotic invariant features, which captures the characteristics of the trajectories. Consequently, a SVM is employed to learn and recognize the human actions using the computed motion features. We performed intensive experiments on multiple benchmark datasets, and obtained promising results. Finally, we consider a more challenging problem referred to as complex event recognition, where the activities of interest are complex and unconstrained. This problem typically pose significant challenges because it involves videos of highly variable content, noise, length, frame size . . . etc. In this extremely challenging task, high-level features have recently shown a promising direction as in [53, 129], where core low-level events referred to as concepts are annotated and modelled using a portion of the training data, then each event is described using its content of these concepts. However, because of the complex nature of the videos, both the concept models and the corresponding high-level features are significantly noisy. In order to address this problem, we propose a novel low-rank formulation, which combines the precisely annotated videos used to train the concepts, with the rich high-level features. Our approach finds a new representation for each event, which is not only low-rank, but also constrained to adhere to the concept annotation, thus suppressing the noise, and maintaining a consistent occurrence of the concepts in each event. Extensive experiments on large scale real world dataset TRECVID Multimedia Event Detection 2011 and 2012 demonstrate that our approach consistently improves the discriminativity of the high-level features by a significant margin.

[pt] Umas das etapas mais complexas e críticas durante a construção de poços de petróleo é o processo de cimentação primária, definido como: O processo de instalação de cimento no anular entre o revestimento e a formação exposta ao poço [1]. Ela fornece isolamento zonal permanente para evitar contaminação ou migração de fluidos indesejáveis no anular, protege o revestimento da ocorrência de corrosão e fornece estabilidade hidráulica e mecânica para o revestimento ao longo da vida produtiva do poço de petróleo. Neste processo, ainda na etapa de perfuração, irregularidades na seção transversal (washouts) podem ser geradas como resultado de diversos colapsos parciais da seção do poço aberto em decorrência da presença de rochas pouco consolidadas da formação. Uma operação de cimentação primária bem sucedida dependerá se o sistema de fluidos espaçadores e a pasta de cimento deslocam de forma adequada e completa o fluido de perfuração do anular e washouts. Motivados por esse problema industrial, a presente dissertação usa um simulador numérico DNS (Direct Numerical Simulation) com o objetivo de analisar sistematicamente o comportamento hidrodinâmico e calcular a eficiência do deslocamento entre dois fluidos newtonianos miscíveis através de um anular contendo uma expansão seguida de uma contração abrupta. Investigamos como diferentes viscosidades e densidades dos fluidos, um injetado e outro deslocado, a miscibilidade entre eles, a taxa de injeção e a dimensão do washout retangular afetam o escoamento bifásico. Consideramos uma geometria axissimétrica durante processos de deslocamentos verticais, e as equações governantes são resolvidas em coordenadas cilíndricas, permitindo investigar diferentes aberturas anulares. Devido à miscibilidade entre os fluidos, nossos resultados preveem eficiências de deslocamento altíssimas, próximas a 100 Fluidos não-newtonianos são usualmente utilizados durante o processo industrial de cimentação primária de poços de petróleo. Contudo, as altas eficiências de deslocamentos encontradas em nossos resultados motivam estudos futuros sobre a influência da miscibilidade em deslocamentos de fluidos complexos. Estes resultados também motivam o uso de fluidos espaçadores para tentar controlar as propriedades de interface. Desta forma, é possível que a utilização de sistemas que se aproximem das condições reológicas e hidrodinâmicas de deslocamento entre fluidos newtonianos miscíveis poderá contribuir para um aumento da eficiência de deslocamento e, consequentemente, otimizar o processo de deslocamento de sistemas de fluidos, tendo em vista melhorias na integridade da cimentação de poços de petróleo. / [en] One of the most complex and critical stages during the construction ofoil wells is the primary cementing process, defined as the process of installingcement in the annulus between the casing and the exposed formation to thewell [1]. Primary cementing provides permanent zonal isolation to preventcontamination or migration of unwanted fluids in the annulus, protects thecasing from corrosion, and provides hydraulic and mechanical stability forthe casing throughout the productive life of the oil well. In this process,during the drilling stage, irregularities in the cross-section (washouts) can begenerated because of various partial collapses of the open wellbore section dueto the presence of poorly consolidated rocks in the formation. A successfulprimary cementing operation will depend on whether the spacer fluid systemand cement slurry adequately and completely displace the drilling fluid fromthe annulus and washouts.Motivated by this industrial problem, the present dissertation uses aDirect Numerical Simulation (DNS) numerical simulator to systematically analyze the hydrodynamic behavior and calculate the displacement efficiency between two miscible newtonian fluids through an annulus containing an expansion followed by an abrupt contraction. We investigate how different viscositiesand densities of the injected and displaced fluids, their miscibility, injectionrate, and the dimension of the rectangular washout affect the two-phase flow.We consider an axisymmetric geometry during vertical displacement processes,and the governing equations are solved in cylindrical coordinates, allowing theinvestigation of different annular clearances. Due to the miscibility betweenthe fluids, our results predict very high displacement efficiencies, close to 100Non-newtonian fluids are commonly used during the industrial processof primary cementing of oil wells. However, the high displacement efficienciesfound in our results motivate further studies on the influence of miscibilityin displacements of complex fluids. These results also encourage the use ofspacer fluids attempting to control the interfacial properties. Therefore, the useof systems that approximate the rheological and hydrodynamic conditions ofdisplacement between miscible newtonian fluids may contribute to an increasein displacement efficiency and, consequently, optimize the displacement processof fluid systems, aiming at improvements in the integrity of well cementing.

On considère le problème d’advection-diffusion stationnaire v(∇u, ∇v)+( β•∇u, v) = (f, v) et non stationnaire d/dt (u(t), v) + v(∇u, ∇v)+( β•∇u, v) = (g(t), v), ainsi que le problème d’advection (β•∇u, v) = (f, v) sur un domaine polygonal borné du plan. Le terme de diffusion est approché par des éléments de Crouzeix Raviart et le terme de convection par une méthode upwind sur des volumes barycentriques finis avec un maillage triangulaire. Pour le problème stationnaire d’advection-diffusion, la L²-stabilité (c’est-à-dire indépendante du coefficient de diffusion v) est démontrée pour la solution du problème approché obtenue par cette méthode d’éléments finis et de volumes finis. Pour cela une condition sur la géométrie doit être satisfaite. Des exemples de maillages sont donnés. Toujours avec cette condition géométrique sur le maillage, une inégalité de stabilité (où la discrétisation en temps n’est pas couplée à une condition sur la finesse du maillage) est obtenue pour le cas non-stationnaire. La discrétisation en temps y est faite par un schéma d’Euler implicite. Une majoration de l’erreur, proportionnelle au pas en temps et à la finesse du maillage, est ensuite proposée et exprimée explicitement en fonction des données du problème. Pour le problème d’advection, une approche utilisant la théorie des graphes est utilisée pour obtenir l’existence et l’unicité de la solution, ainsi que le résultat de stabilité. Comme pour la stabilité du problème d’advection-diffusion, une condition géométrique - qui est équivalente pour les points intérieurs du maillage à celle du problème d’advection-diffusion - est nécessaire. / We consider the stationary linear convection-diffusion equation v(∇u, ∇v)+( β•∇u, v) = (f, v), the time dependent d/dt (u(t), v) + v(∇u,∇v)+( β•∇u, v)= (g(t), v) equation and the linear advection equation (β•∇u, v) = (f, v) on a two dimensional bounded polygonal domain. The diffusion term is discretized by Crouzeix-Raviart piecewise linear finite elements, and the convection term by upwind barycentric finite volumes on a triangular grid. For the stationary convection-diffusion problem, L²-stability (i.e. independent of the diffusion coefficient v) is proven for the approximate solution obtained by this combined finite-element finite-volume method. This result holds if the underlying grid satisfies a condition that is fulfilled, for example, by some structured meshes. Using again this condition on the grid, stability is shown for the time dependent convection-diffusion equation (without any link between mesh size and time step). An implicit Euler approach is used for the time discretization. It is shown that the error associated with this scheme decays linearly with the mesh size and the time step. This result holds without any link between mesh size and time step. The dependence of the corresponding error bound on the diffusion coefficient is completely explicit. For the stationary advection equation, an approach using graph theory is used to obtain existence, uniqueness and stability. As in the stationary linear convection-diffusion equation, the underlying grid must satisfy some geometric condition. / Gegenstand der Arbeit ist die zweidimensionale stationäre Konvektion-Diffusionsgleichung v(∇u, ∇v)+( β•∇u, v) = (f, v), die zeitabhängige Konvektion-Diffusionsgleichung d/dt (u(t), v) + v(∇u,∇v)+( β•∇u, v)= (g(t), v), sowie die Konvektionsgleichung (β•∇u, v) = (f, v). Der Diffusionsterm ist diskretisiert mittels Crouzeix-Raviart stückweise lineare Finite Elemente. Das Gebiet ist in Dreiecke unterteilt und der Konvektionsterm ist mittels einer upwind Methode auf Baryzentrische Finite Volumenelemente definiert. Für die stationäre Konvektion-Diffusionsgleichung, wird (d.h. von v unabhängige) L²-Stabilität der numerischen Lösung bewiesen. Voraussetzung dafür, ist die Erfüllung gewisser geometrischer Bedingungen an die Unterteilung des Gebiets. Beispiele von Unterteilungen die diese Bedingungen erfüllen, werden gegeben. Wieder an dieser geometrischen Bedingung geknüpft, wird Stabilität (d.h. die Zeitdiskretisierung ist entkoppelt von der Netzweite) für die zeitabhängige Konvektion-Diffusionsgleichung, bewiesen. Für die Zeitableitung wird dabei eine Implizite Euler Diskretisierung verwendet. Eine obere Schranke für den Diskretisierungsfehler, proportional zum Zeitdiskretisierungsparameter und zur Netzfeinheit, ausgedrückt als Funktion der Daten der Differenzialgleichung, wird gezeigt. Für die Konvektionsgleichung wird ein graphentheoretischer Zugang verwendet, der es ermöglicht Existenz, Eindeutigkeit und Stabilität, zu bekommen. Für die Stabilität, werden ähnliche geometrische Bedingungen an die Unterteilung des Gebiets gestellt, wie beim stationären Konvektion-Diffusionsproblem.

In dieser Arbeit betrachten wir das Strömungs-Diffusions-(Reaktions)-Problem für passive Markerteilchen, die in zweidimensionalen laminaren Strömungsmustern mit geringem thermischem Rauschen gelöst sind. Der deterministische Fluss umfasst Zellen in Form von Quadraten oder Katzenaugen. In ihnen tritt Rotationsbewegung auf. Einige der Strömungen bestehen aus wellenförmigen Bereichen mit gerader Vorwärtsbewegung. Alle Systeme sind entweder periodisch oder durch Wände begrenzt. Eine untersuchte Familie von Strömungen interpoliert kontinuierlich zwischen Reihen von Wirbeln und Scherflüssen. Wir analysieren zahlreiche numerische Simulationen, die bisherige theoretische Vorhersagen bestätigen und neue Phänomene offenbaren. Ohne Rauschen sind die Teilchen in einzelnen Bestandteilen des Flusses für immer gefangen. Durch Hinzufügen von schwachem thermischen Rauschen wird die normale Diffusion für lange Zeiten stark verstärkt und führt zu verschiedenen Diffusionsarten für mittlere Zeiten. Mit Continuous-Time-Random-Walk-Modellen leiten wir analytische Ausdrücke in Übereinstimmung mit den numerischen Ergebnissen her, die je nach Parametern, Anfangsbedingungen und Alterungszeiten von subdiffusiver bis superballistischer anomaler Diffusion für mittlere Zeiten reichen. Wir sehen deutlich, dass einige der früheren Vorhersagen nur für Teilchen gelten, die an der Separatrix des Flusses starten - der einzige Fall, der in der Vergangenheit ausführlich betrachtet wurde - und dass das System zu vollkommen anderem Verhalten in anderen Situationen führen kann, einschließlich einem Schwingenden beim Start im Zentrum einesWirbels nach einer gewissen Alterungszeit. Darüber hinaus enthüllen die Simulationen, dass Teilchenreaktionen dort häufiger auftreten, wo sich die Geschwindigkeit der Strömung stark ändert, was dazu führt, dass langsame Teilchen von schnelleren getroffen werden, die ihnen folgen. Die umfangreichen numerischen Simulationen, die für diese Arbeit durchgeführt wurden, mussten jetzt durchgeführt werden, da wir die Rechenleistung dafür besitzen. / In this thesis, we consider the advection-diffusion-(reaction) problem for passive tracer particles suspended in two-dimensional laminar flow patterns with small thermal noise. The deterministic flow comprises cells in the shape of either squares or cat’s eyes. Rotational motion occurs inside them. Some of the flows consist of sinusoidal regions of straight forward motion. All systems are either periodic or are bounded by walls. One examined family of flows continuously interpolates between arrays of eddies and shear flows. We analyse extensive numerical simulations, which confirm previous theoretical predictions as well as reveal new phenomena. Without noise, particles are trapped forever in single building blocks of the flow. Adding small thermal noise, leads to largely enhanced normal diffusion for long times and several kinds of diffusion for intermediate times. Using continuous time random walk models, we derive analytical expressions in accordance with numerical results, ranging from subdiffusive to superballistic anomalous diffusion for intermediate times depending on parameters, initial conditions and aging time. We clearly see, that some of the previous predictions are only true for particles starting at the separatrix of the flow - the only case considered in depth in the past - and that the system might show a vastly different behavior in other situations, including an oscillatory one, when starting in the center of an eddy after a certain aging time. Furthermore, simulations reveal that particle reactions occur more frequently at positions where the velocity of the flow changes the most, resulting in slow particles being hit by faster ones following them. The extensive numerical simulations performed for this thesis had to be done now that we have the computational means to do so. Machines are powerful tools in order to gain a deeper and more detailed insight into the dynamics of many complicated dynamical and stochastic systems.

O objetivo do trabalho é aplicar um modelo numérico que simule a propagação da pluma de contaminação derivada de chorume, como ferramenta numérica que permita uma análise dos parâmetros envolvidos no sistema, tendo como caso de estudo o lixão de São Pedro da Aldeia, levando em conta: o mecanismo de perda de massa durante o transporte e a variabilidade estatística dos principais parâmetros de entrada (input) no modelo numérico. A perda de massa durante o fluxo pode ocorrer tanto por sorção, como por decaimento microbiológico, será estimada através da interpretação probabilística de dados experimentais obtidos no lixão de São Pedro da Aldeia. Os resultados obtidos pela modelagem numérica serão comparados com os resultados experimentais de concentração do íon cloreto obtidos por medições em campo. Os resultados experimentais foram obtidos considerando distâncias crescentes em relação ao lixão, para oito intervalos de tempo distintos, permitindo, assim, comparar os resultados alcançados pela modelagem numérica com dados experimentais reais, obtidos em função do tempo e da distância ao lixão. Para a modelagem numérica, foi utilizada a equação diferencial do transporte advectivo-dispersivo, levando em conta um termo de perda de massa estimado a partir de interpretações estatísticas dos resultados experimentais obtidos. A solução da equação diferencial do fluxo advectivo dispersivo foi analisada pelo método dos volumes finitos, que se mostra adequado para problemas de avanço de frentes de saturação e/ou de contaminação. Para a modelagem probabilística, foi utilizado o Método de Rosenblueth (ou método das estimativas pontuais), considerando-se, como variáveis aleatórias mais relevantes: a condutividade hidráulica k, a concentração iônica inicial Co e o decaimento da concentração iônica em função da distância (x) ao lixão expresso por um termo bo em (mg/l/m). A interpretação estatística das concentrações medidas experimentalmente mostrou uma faixa de variabilidade muito bem representada pela faixa de variação obtida pela conjugação da solução da equação de transporte de contaminantes (pelo método dos volumes finitos) com análise de variabilidade por formulação FOSM (pelo método das estimativas pontuais de Rosenblueth). Os resultados obtidos indicam que, embora com a contaminação crescente, podemos avaliar com poucos dados a contaminação ao longo do tempo. As simulações indicaram coerência com o resultado obtido experimentalmente, validando assim o modelo utilizado. Entende-se que o trabalho conseguiu demonstrar a aplicabilidade e relativa praticidade no que diz respeito ao estudo do fluxo de contaminantes em meios saturados. Os resultados se mostraram satisfatórios no que tange ao entendimento das questões abordadas. / The objective of this work is to apply a numerical model that simulates the propagation of the contamination plume stemmed from leachate as a numerical tool that allows an analysis of the parameters involved in the system, having the landfill of São Pedro da Aldeia as study case. The model takes into account the loss of contaminant mass mechanism during transport and the statistical variability, in the numerical model, of the main input variables. The loss of mass during the flow, which can occur either in the form of sorption or as a microbiological decay, will be assed by probabilistic interpretation of experimental data. The results obtained by the numerical modeling will be compared to the experimental results of the ionic chloride concentration obtained by field mediations. The experimental results were obtained taking into account the growing distances in relation to the landfill, for eight different time intervals, allowing, this way, the comparison of the results reached by the numerical modeling with the real experimental data results, obtained due to the time and the distance of the landfill. For the numerical modeling, the differential equation of the advective-dispersive transport was used, taking into account a term of loss of mass estimated from statistical interpretations of the experimental results. The differential equation solution of the advective-dispersive flow was obtained by the finite volumes method, which was proved adequate to problems concerning the propagation of contaminating plumes in saturated soils. For the probabilistic modeling, the Rosenblueth Method (or point estimate method) was used, considering, as more relevant aleatory variables, the hydraulic conductivity k, the initial ionic concentration and the rate of concentration decay. The statistical interpretation of the concentrations experimentally measured showed one variability band well represented by the variation band obtained by the conjunction of the contaminant transport equation solution (finite volumes method) with the variability analysis of the FOSM formularization (Rosenblueths point estimate method). The obtained results indicate that even with the increasing contamination it is possible, with little data, to evaluate the contamination throughout time. The simulations indicated coherence with the experimentally obtained results validating, this way, the used model. The work demonstrates the applicability and the relative practicity concerning the study of the contaminant flow in saturated environments. Considering the understanding of the boarded questions, the results seemed satisfactory.

O objetivo do trabalho é aplicar um modelo numérico que simule a propagação da pluma de contaminação derivada de chorume, como ferramenta numérica que permita uma análise dos parâmetros envolvidos no sistema, tendo como caso de estudo o lixão de São Pedro da Aldeia, levando em conta: o mecanismo de perda de massa durante o transporte e a variabilidade estatística dos principais parâmetros de entrada (input) no modelo numérico. A perda de massa durante o fluxo pode ocorrer tanto por sorção, como por decaimento microbiológico, será estimada através da interpretação probabilística de dados experimentais obtidos no lixão de São Pedro da Aldeia. Os resultados obtidos pela modelagem numérica serão comparados com os resultados experimentais de concentração do íon cloreto obtidos por medições em campo. Os resultados experimentais foram obtidos considerando distâncias crescentes em relação ao lixão, para oito intervalos de tempo distintos, permitindo, assim, comparar os resultados alcançados pela modelagem numérica com dados experimentais reais, obtidos em função do tempo e da distância ao lixão. Para a modelagem numérica, foi utilizada a equação diferencial do transporte advectivo-dispersivo, levando em conta um termo de perda de massa estimado a partir de interpretações estatísticas dos resultados experimentais obtidos. A solução da equação diferencial do fluxo advectivo dispersivo foi analisada pelo método dos volumes finitos, que se mostra adequado para problemas de avanço de frentes de saturação e/ou de contaminação. Para a modelagem probabilística, foi utilizado o Método de Rosenblueth (ou método das estimativas pontuais), considerando-se, como variáveis aleatórias mais relevantes: a condutividade hidráulica k, a concentração iônica inicial Co e o decaimento da concentração iônica em função da distância (x) ao lixão expresso por um termo bo em (mg/l/m). A interpretação estatística das concentrações medidas experimentalmente mostrou uma faixa de variabilidade muito bem representada pela faixa de variação obtida pela conjugação da solução da equação de transporte de contaminantes (pelo método dos volumes finitos) com análise de variabilidade por formulação FOSM (pelo método das estimativas pontuais de Rosenblueth). Os resultados obtidos indicam que, embora com a contaminação crescente, podemos avaliar com poucos dados a contaminação ao longo do tempo. As simulações indicaram coerência com o resultado obtido experimentalmente, validando assim o modelo utilizado. Entende-se que o trabalho conseguiu demonstrar a aplicabilidade e relativa praticidade no que diz respeito ao estudo do fluxo de contaminantes em meios saturados. Os resultados se mostraram satisfatórios no que tange ao entendimento das questões abordadas. / The objective of this work is to apply a numerical model that simulates the propagation of the contamination plume stemmed from leachate as a numerical tool that allows an analysis of the parameters involved in the system, having the landfill of São Pedro da Aldeia as study case. The model takes into account the loss of contaminant mass mechanism during transport and the statistical variability, in the numerical model, of the main input variables. The loss of mass during the flow, which can occur either in the form of sorption or as a microbiological decay, will be assed by probabilistic interpretation of experimental data. The results obtained by the numerical modeling will be compared to the experimental results of the ionic chloride concentration obtained by field mediations. The experimental results were obtained taking into account the growing distances in relation to the landfill, for eight different time intervals, allowing, this way, the comparison of the results reached by the numerical modeling with the real experimental data results, obtained due to the time and the distance of the landfill. For the numerical modeling, the differential equation of the advective-dispersive transport was used, taking into account a term of loss of mass estimated from statistical interpretations of the experimental results. The differential equation solution of the advective-dispersive flow was obtained by the finite volumes method, which was proved adequate to problems concerning the propagation of contaminating plumes in saturated soils. For the probabilistic modeling, the Rosenblueth Method (or point estimate method) was used, considering, as more relevant aleatory variables, the hydraulic conductivity k, the initial ionic concentration and the rate of concentration decay. The statistical interpretation of the concentrations experimentally measured showed one variability band well represented by the variation band obtained by the conjunction of the contaminant transport equation solution (finite volumes method) with the variability analysis of the FOSM formularization (Rosenblueths point estimate method). The obtained results indicate that even with the increasing contamination it is possible, with little data, to evaluate the contamination throughout time. The simulations indicated coherence with the experimentally obtained results validating, this way, the used model. The work demonstrates the applicability and the relative practicity concerning the study of the contaminant flow in saturated environments. Considering the understanding of the boarded questions, the results seemed satisfactory.

[pt] Nesta tese, a teoria das hastes de Cosserat é revisitada e aplicada à dinâmica de coluna de perfuração. O objetivo é estudar o comportamento dessas estruturas dentro de poços de petróleo curvos. Para atingir este objetivo, um modelo estrutural determinístico é construído onde as tubos de perfuração (drill-pipes) e o conjunto de fundo (bottom hole assembly) são considerados como uma estrutura unidimensional de Cosserat. Em seguida, é desenvolvida uma estratégia para tratar o contato lateral em poços com configuração curvilínea. Depois disso, o problema de contorno livre é tratado mediante uma estratégia que considera como a condição de borda evolui à medida que a estrutura de perfuração avança. Isto é feito mediante uma formulação de interação broca-rocha que deve considerar a dinâmica de corte. Para isso, uma equação extra, de advecção, é resolvida junto com as equações de movimento de Cosserat. Em seguida, alguns casos de aplicação são apresentados. Numa primeira instancia, alguns elementos do problema são avaliados separadamente. Seguidamente, eles são integrados e analisados de forma conjunta. Por exemplo, primeiramente uma coluna de perfuração sem contato de fundo (off-bottom) é simulada, ou seja, sem contato broca-rocha, para estudar o comportamento e a implementação da estratégia para o contato lateral. Aqui também são calibrados alguns dos parâmetros do modelo de atrito. Em seguida, a estratégia para contabilizar o corte na rocha é implementada em um modelo 2-DOF de baixa dimensão e em um semi-discreto onde a dinâmica de torção é modelada como uma equação de onda. Os resultados mostram que o uso de abordagens contínuas resulta mais apropriade que aquelas onde se utilizam modelos de baixa dimensãom, particularmente quando são consideradas colunas longas, e quando há interesse em analisar não apenas o comportamento da broca, mas também o comportamento do sistema mecânico ao longo dos tubos de perfuração. Isso é reforçado por outro exemplo onde a dinâmica de corte é combinada com a formulação de Cosserat. Observações semelhantes do ponto de vista qualitativo são encontradas. Resumindo os resultados obtidos, as diferenças nas previsões dadas pelos modelos de baixa dimensão e o de unidimensional de Cosserat justificam o desenvolvimento e aplicação da abordagem com esta formulação em estruturas de perfuração. Finalmente, a modo de introduzir outro aspecto importante em colunas de perfuração e que pode ser uma linha de pesquisa para continuar o trabalho, a variabilidades presente em elementos como rocha, inclui-se um caso de aplicação considerando um poço horizontal e um campo estocástico de atrito. / [en] In this thesis, the theory of Cosserat rods is applied to the dynamics of drill-strings. The main objective is to evaluate the behaviour of these strings when they move within curved wells. To achieve this goal, a deterministic structural model is constructed, where the drill-pipes and the bottom hole assembly are taken as a Cosserat rod. Next, a strategy to deal with the lateral contact in curved well configurations is developed. After that, the free boundary problem is assessed: while drilling, the boundary changes due to cutting, modifying the position of the soil and, consequently, changing the bit-rock interaction forces. For this reason, a bit-rock model that can account for the cutting dynamics is adopted, in which an extra advection equation is solved together with the equations of motion of the Cosserat rod. Next, application cases are provided. First, some effects included in the model are tested in isolation, such as the lateral friction, the lateral contact, and the cutting. After that, they are all combined. In the first analysis, an off-bottom string is simulated, i.e. without contact at the bit. This allows testing the formulation associated with the lateral contact. Also, the calibration of the lateral friction parameters is made. Following that, the strategy to account for the cutting at the bit is implemented in a low-dimensional 2-DOF model, and in a semi-discrete model with a continuous wave equation for the torsional dynamics. The results show that the use of continuous approaches is more appropriate than low-dimensional models. Especially when long columns are considered, and when there is interest in understanding not only the behaviour at the bit but also along drill-pipes. This finding is reinforced by another application where the cutting dynamics are combined with the Cosserat rod formulation. Again, similar observations from a qualitative point of view are found. Overall, the differences in the results between the lumped low-dimensional models and the continuous Cosserat rod justify the development and application of the Cosserat approach to drilling structures. Finally, an introductory stochastic analysis concerning the variability of the rock is presented as an introduction to a future line of research, where stochasticity is included. / [es] En esta tesis, la teoria de Cosserat para elementos unidimensionales es revisitada y aplicada a la simulación de columnas de perforación. El objetivo es estudiar el comportamiento de estas estructuras en pozos de geometría curva. Para alcanzar este objetivo se construye un modelo determinístico. En este modelo, los caños de perforación (drill-pipes) y el conjunto de fondo (bottom hole assembly) son modelados como una estructura unidimensional de Cosserat. Seguidamente, una estrategia para tratar con el contacto lateral en pozos curvos es desarrollada. Luego, el problema de frontera libre es estudiado: durante la perforación, la condición de borde cambia debido al cambio del perfil altimétrico del terreno, alterando su posición y consecuentemente las fuerzas asociadas a la interacción broca-roca. Por esta razón, se decide utilizar un modelo de interacción broca-roca que tiene en cuenta la dinámica del corte. En este abordaje una ecuación extra, la ecuación de advección, es resuelta en forma acoplada con las ecuaciones del movimiento de la estructura de Cosserat. Algunos ejemplos de aplicación son presentados. En una primera instancia, algunos de los elementos del problema son estudiados en forma aislada. Luego combinados en un modelo completo. Por ejemplo, el caso de una columna sin contacto de fondo (off-bottom) es tratado para evaluar el comportamiento y la implementación de la estrategia mencionada para detectar el contacto lateral. Además, se efectúa la calibración de alguno de los parámetros relacionados con la fricción lateral. Luego, la estrategia para considerar el corte en la punta es implementada en un modelo de 2-DOF, y en otro semi-discreto donde se considera un modelo de ecuación de onda para la dinámica torsional. Los resultados muestran que el uso de formulaciones continuas es más apropiado que aquellas formulaciones donde se utilizan modelos de dimensiones reducidas, particularmente cuando se estudia columnas largas donde el interés se centra en entender no solo el comportamiento de la broca sino también a lo largo de la tubería. Este resultado es reforzado por otro caso de aplicación en donde se combina la dinámica de corte con un modelo de Cosserat. Observaciones similares son vistas en el comportamiento cualitativo de la solución. En resumen, las diferencias observadas en los diferentes ejemplos de aplicación entre los modelos de dimensiones reducidas y el modelo continuo de Cosserat justifican el desarrollo y la aplicación de la teoría de Cosserat a estructuras de perforación. Finalmente, dado que uno de los objetivos planteados también es considerar la variabilidad en algunos elementos como ser las propiedades de la roca, un caso de aplicación considerando un pozo horizontal es mostrado.

This thesis presents the development and evaluation of a 2D simulation model designed to predict the spread and behavior of forest fires, with a specific focus on Swedish forest ecosystems. Despite the model's simplicity and inherent limitations due to assumptions such as homogeneity in fuel distribution and the exclusion of topographical influences, the simulations yield remarkably accurate predictions of fire spread and intensity. The model integrates basic meteorological data (wind speed and direction, temperature) and uses a discretized spatial approach to simulate the dynamics of forest fires. Initial results suggest that even with minimal input variables and broad assumptions, the model offers significant predictive capabilities, highlighting potential areas for future refinement. Key aspects such as the interaction between conduction and advection terms, the role of water vaporization in fire dynamics, and the influence of wind on fire propagation are discussed. The findings encourage further development of the model, aiming at incorporating more complex variables such as topography and more forest fuels, potentially enhancing its utility in real-time fire management and decision-making processes.

A high-order central essentially non-oscillatory (CENO) finite-volume scheme in combination with a block-based adaptive mesh refinement (AMR) algorithm is proposed for solution of hyperbolic and elliptic systems of conservation laws on body- fitted multi-block mesh. The spatial discretization of the hyperbolic (inviscid) terms is based on a hybrid solution reconstruction procedure that combines an unlimited high-order k-exact least-squares reconstruction technique following from a fixed central stencil with a monotonicity preserving limited piecewise linear reconstruction algorithm. The limited reconstruction is applied to computational cells with under-resolved solution content and the unlimited k-exact reconstruction procedure is used for cells in which the solution is fully resolved. Switching in the hybrid procedure is determined by a solution smoothness indicator. The hybrid approach avoids the complexity associated with other ENO schemes that require reconstruction on multiple stencils and therefore, would seem very well suited for extension to unstructured meshes. The high-order elliptic (viscous) fluxes are computed based on a k-order accurate average gradient derived from a (k+1)-order accurate reconstruction. A novel h-refinement criterion based on the solution smoothness indicator is used to direct the steady and unsteady refinement of the AMR mesh. The predictive capabilities of the proposed high-order AMR scheme are demonstrated for the Euler and Navier-Stokes equations governing two-dimensional compressible gaseous flows as well as for advection-diffusion problems characterized by the full range of Peclet numbers, Pe. The ability of the scheme to accurately represent solutions with smooth extrema and yet robustly handle under-resolved and/or non-smooth solution content (i.e., shocks and other discontinuities) is shown for a range of problems. Moreover, the ability to perform mesh refinement in regions of smooth but under-resolved and/or non-smooth solution content to achieve the desired resolution is also demonstrated.