Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2015-01-13T10:55:38Z No. of bitstreams: 2 license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Dissertação - Gustavo Batista de Castro Souza - 2014.pdf: 11760996 bytes, checksum: 8245af285d79ff9e8079bafddb72e690 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2015-01-13T10:56:54Z (GMT) No. of bitstreams: 2 license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Dissertação - Gustavo Batista de Castro Souza - 2014.pdf: 11760996 bytes, checksum: 8245af285d79ff9e8079bafddb72e690 (MD5) / Made available in DSpace on 2015-01-13T10:56:54Z (GMT). No. of bitstreams: 2 license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Dissertação - Gustavo Batista de Castro Souza - 2014.pdf: 11760996 bytes, checksum: 8245af285d79ff9e8079bafddb72e690 (MD5) Previous issue date: 2014-07-17 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Smart Grids systems have become widespread around the world. The RF mesh communication systems have contributed to make power systems smarter and reliable with implementation of Distributed Automation and Demand Response technologies. This work presents a methodology for positioning of GPRS concentrators in a energy meter ZigBee mesh network in order to attain the average network delay, thus aiming to improve the performance of the communication service. The proposed algorithm determines the amount and placement of concentrators using Integer Linear Programming and a Queuing Model for the Mesh Network. The solutions given by the proposed algorithm are validated by verifying the network performance through computer simulations based on real network scenarios. / Smart Grids tem se difundido em todo o mundo. Sistemas de comunicação RF Mesh (em malha) tem contribuído para deixar sistemas de potência mais inteligentes e confiáveis com a implantação de tecnolgias de Automação da Distribuição e Resposta à Demanda. Este trabalho apresenta um metodologia de posicionamento de concentradores GPRS em uma rede ZigBee mesh de medidores de energia elétrica com o objetivo de limitar o delay médio da rede e assim otimizar o desempenho do serviço de comunicação. O algoritmo proposto determina a quantidade e a localização de concentradores utilizando Programação Linear Inteira e um Modelo de Filas para Redes Mesh. As soluções obtidas pelo algoritmo proposto são validadas verificando o desempenho da rede através de simulações computacionais baseadas em cenários reais de redes.

O interesse em escoamento bifásico líquido-líquido aumentou recentemente, em especial devido às grandes perdas de energia envolvidas no transporte de petróleo, onde comumente uma mistura bifásica óleo-água é deslocada ao longo de grandes distancias. Embora este tipo de escoamento seja comum na indústria, não existem tantos trabalhos na literatura quanto os relacionados ao escoamento gás-líquido. Alguns estudos sobre escoamentos óleo-água reportam uma redução de atrito em dispersões e emulsões em regime turbulento sem adição de qualquer tipo de substancia química, mas a física por trás do fenômeno ainda não é bem compreendida. Neste trabalho, foi estudado o padrão de escoamento disperso óleo-água em tubulação horizontal, visando a obtenção de novos dados experimentais e um melhor entendimento do fenômeno de redução de atrito. Uma série de parâmetros considerados importantes para a caracterização do escoamento foi investigada: queda de pressão, fração volumétrica e sub-padrões de escoamento disperso. Apresentam-se dados de distribuição das fases e fração volumétrica in situ, obtidos através de um moderno sensor intrusivo, do tipo wire-mesh, baseado em medidas de capacitância (permissividade). Câmera de alta velocidade e técnica das válvulas de fechamento rápido foram utilizadas para validar os sinais do sensor. Um modelo prospectivo simplificado foi desenvolvido como uma tentativa de explicar a ocorrência do fenômeno de redução de atrito no padrão disperso óleo-água estudado. O modelo sugere que a presença de uma fina película de água rente a parede hidrofílica/oleofóbica do tubo poderia explicar a diminuição no gradiente de pressão bifásico observada. / The interest in two-phase liquid-liquid flow has increased recently mainly due to the petroleum industry where oil and water are often produced and transported together for long distances and the significant frictional pressure gradient involved. Liquid-liquid flows are present in a wide range of industrial processes; however, they have not been studied as intensively as gas-liquid flows. Drag reduction phenomenon in oil-water flows without the addition of any drag reduction agent has been detected in previous works, but the physics behind the phenomenon is yet not well understood. The aim of the research was to study the dispersed oil-water flow pattern in a horizontal pipe in order to better understand the phenomenon and the obtaining of new experimental data of oil-water dispersed flows. Important issues related to oil-water pipe flow were investigated: pressure drop, volume fraction and flow patterns. Phase distribution and holdup data were obtained by a new wire-mesh sensor based on capacitance (permittivity) measurements. A high-speed video camera and the Quick-Closing-Valves technique were used to compare and validate the signals of the wire-mesh sensor. A simplified mathematical model was proposed to explain the drag reduction phenomenon in the oil-water dispersed flow studied. The model suggests that the presence of a thin water film between the homogenously dispersed flow and the pipe wall could explain the observed decreases of the two-phase pressure gradient.

A ocorrência frequente de escoamentos multifásicos em tubulações tem motivado um grande interesse acadêmico nas últimas décadas. O caso particular de escoamentos líquidolíquido é geralmente encontrado na indústria do petróleo, onde uma série de aplicações envolve escoamentos óleo-água, tais como a produção de petróleo e seu transporte. No entanto, ele não tem recebido a mesma atenção quando comparado com escoamentos gás-líquido. Ainda não existe uma explicação física razoável para um número significativo de fenômenos observados em escoamento óleo-água, como o fenômeno de redução de atrito, observado em escoamento disperso, e a estrutura interfacial ondulada em escoamento estratificado. Os escoamentos trifásicos têm sido ainda menos estudados. Há técnicas de medição, desenvolvidas para escoamento gás-líquido, que supostamente são adaptáveis aos escoamentos líquido-líquido, mas várias delas ainda precisam de validação adequada. O sensor wire-mesh, um método híbrido baseado na medição de resistência ou capacitância, que combina medição local intrusiva da fração de fase e imagens tomográficas transversais, oferece uma boa resolução espacial e alta resolução temporal em comparação com outras técnicas atuais. Porém, a literatura existente em aplicações do sensor wire-mesh cobre quase apenas o escoamento gás-líquido e, até agora, não é possível avaliar o limite de viscosidade do fluido para a sua aplicação. Neste contexto, este projeto propõe o estudo de aspectos importantes da fenomenologia de escoamentos líquido-líquido e líquido-líquido-gás com o auxílio da tecnologia wire-mesh. O principal objetivo prático é a validação da técnica wire-mesh como ferramenta de referência para o desenvolvimento de instrumentação para aplicações em campos petrolíferos, com especial atenção devotada ao efeito da viscosidade do óleo sobre a confiabilidade da medição e à extensão da tecnologia para lidar com escoamentos óleo-água-gás. / The frequent occurrence of multiphase flows in pipes has motivated a great research interest over the last decades. The particular case of liquid-liquid flow is commonly encountered in the petroleum industry, where a number of applications involve oil-water flow such as crude oil production and transportation. However, it has not received the same attention when compared to gas-liquid flow. There is no reasonable physical explanation for a significant number of phenomena observed in oil-water flow, as the drag reduction phenomenon observed in dispersed flow and the interfacial wavy structure of stratified flow. Much less has been investigated when it comes to three-phase flow. Several measurement techniques that are supposed to be adaptable to liquid-liquid flow have been proposed recently, but many of them still need proper validation. The wire-mesh sensor, a hybrid impedance-based method that combines intrusive local measurement of phase fraction and tomographic cross-sectional imaging, offers good spatial resolution and high temporal resolution in comparison with other current techniques. However, the existing literature on wire-mesh sensor applications covers almost only the gas-liquid flow and, so far, it is not possible to evaluate the fluid-viscosity limit for its application. In that context, this project proposes the study of important aspects of liquidliquid and liquid-liquid-gas flow phenomenology with the aid of the wire-mesh technology. The main goal is the validation of the wire-mesh technique as a reference tool for the development of instrumentation for oilfield application, with especial attention paid to the effect of oil viscosity on measurement reliability and the extension of the technology for dealing with oil-water-gas flow.

This paper is concerned with mesh refinement techniques for treating elliptic boundary value problems in domains with re- entrant edges and corners, and focuses on numerical experiments. After a section about the model problem and discretization strategies, their realization in the experimental code FEMPS3D is described. For two representative examples the numerically determined error norms are recorded, and various mesh refinement strategies are compared.

We address several problems that arise in a multihop wireless mesh network. First, we study the problem of joint congestion control, routing and MAC layer scheduling. We formulate the problem as an aggregate utility maximization problem and apply duality theory to decompose the problem into two sub-problems, namely, network layer congestion control and routing problem, and MAC layer scheduling problem. Given the link “prices", the source adjusts its rate based on the cost of the least-cost path to the destination, and sends traffic to the destination along the least-cost path, while link scheduling is carried out based on link prices. Optimal link scheduling for a wireless network is known to be NP-hard. We explore the use of a known centralized greedy heuristic, and develop a distributed algorithm that can schedule independent links based on local information. While the link scheduling algorithm above is for a given set of link prices, the solution to our problem depends on the sequence of price vectors generated by the price update algorithm. This leads us to study convergence issues related to the price update algorithm. Next, we develop a practical protocol which maximizes aggregate utility in a wireless mesh network. We simulate our protocol using Qualnet 4.5 and compare the result with a baseline protocol that uses IEEE 802.11 for link scheduling and AODV for routing. Our proposed protocol requires the durations of slots and subslots to be defined. We develop an approach in which given a single cell wireless mesh network using IEEE 802.11 as a reliable message delivery mechanism, one can find upper and lower bounds on the durations of slots. We employ stochastic ordering to compare distributions of random variables and using some properties of stochastic ordering along with the central limit theorem, we devise a way to compute the above mentioned bounds on the durations. In the second part, we shift our focus to model delays incurred by application packets sent over a WLAN. In this section we model the WLAN as a Random Polling System. The packet arrival process at each node i is assumed to be a stationary and independent increment random process with mean ai and second moment a(2)i . The packet lengths at node i are assumed to be i.i.d random variables Pi with finite mean and second moment. Utilizing available results, we obtain expressions for mean packet delay. Extensive simulations are conducted to verify the analytical results.

Adaptive Mesh Refinement (AMR) is a method which dynamically varies the spatio-temporal resolution of localized mesh regions in numerical simulations, based on the strength of the solution features. Due to high resolution discretization of localized regions of interests into rectangular mesh units called patches, AMR provides low cost of computations and high degree of accuracy. General purpose graphics processing units (GPGPUs) with their support for fine-grained parallelism, offer an attractive option for obtaining high performance for AMR applications. The data parallel computations of the finite difference schemes of AMR can be efficiently performed on GPGPUs. This research deals with challenges and develops techniques for efficient executions of AMR applications with uniform and non-uniform patches on GPUs. In the first part of the thesis, we optimize an AMR model with uniform patches. We have developed strategies for continuous online visualization of time evolving data for AMR applications executed on GPUs. In-situ visualization plays an important role for analyzing the time evolving characteristics of the domain structures. Continuous visualization of the output data for various time steps results in better study of the underlying domain and the model used for simulating the domain. We reorder the meshes for computations on the GPU based on the users input related to the subdomain that he wants to visualize. This makes the data available for visualization at a faster rate. We then perform asynchronous executions of the visualization steps and fix-up operations on the coarse meshes on the CPUs while the GPU advances the solution. By performing experiments on Tesla S1070 and Fermi C2070 clusters, we found that our strategies result in up to 60% improvement in response time and 16% improvement in the rate of visualization of frames over the existing strategy of performing fix-ups and visualization at the end of the time steps. The second part of the thesis deals with adaptive strategies for efficient execution of block structured AMR applications with non-uniform patches on GPUs. Most AMR approaches use patches of uniform sizes over regions of interests. Since this leads to over-refinement, some efforts have focused on forming patches of non-uniform dimensions to improve computational efficiency since the dimensions of a patch can be tuned to the geometry of a region of interest. While effective hybrid execution strategies exist for applications with uniform patches, our work considers efficient execution of non-uniform patches with different workloads. Our techniques include a geometric bin-packing method to load balance GPU computations and reduce thread idling, adaptive determination of amount of work to maximize asynchronism between CPU and GPU executions using a knapsack formulation, and scheduling communications for multi-GPU executions. We test our strategies for synthetic inputs as well as for traces from real applications. Our experiments on Tesla S1070 and Fermi C2070 clusters with both single-GPU and multi-GPU executions show that our strategies result in up to 69% improvement in performance over existing strategies. Our bin-packing based load balancing gives performance gains up to 39%, kernel optimizations give an improvement of up to 20%, and our strategies for adaptive asynchronism between CPU-GPU executions give performance improvements of up to 17% over default static asynchronous executions.

En ce qui concerne les problèmes de Dynamique des Fluides Numériques, l’adaptation du maillage est intéressante pour sa capacité à aborder la convergence asymptotique et à obtenir une prévision précise pour des flux complexes à moindre coût. La méthode d’adaptation de maillage anisotrope réduit le nombre de degrés de liberté nécessaires pour atteindre la précision d’une solution donnée, ce qui a un impact positif sur le temps de calcul. De plus, il réduit la dissipation du schéma numérique en tenant compte automatiquement de l'anisotropie des phénomènes physiques à l'intérieur du maillage. Deux approches principales existent dans la littérature. L'adaptation du maillage basée sur les caractéristiques géométriques, qui est principalement déduite d'une estimation de l'erreur d'interpolation utilisant la hessienne du senseur choisi, contrôle l'erreur d'interpolation du capteur sur l'ensemble du domaine de calcul. Une telle approche est facile à mettre en place et a un large éventail d’applications, mais elle ne prend pas en compte l’EDP considérée utilisée pour résoudre le problème. D'autre part, l'adaptation de maillage orientée fonctionnelle, qui se concentre sur une fonctionnelle scalaire, prend en compte à la fois la solution et l'EDP dans l'estimation d'erreur grâce à l'état adjoint. Mais, la conception de cette estimation d'erreur est beaucoup plus compliquée. Cette thèse présente les résultats obtenus avec différentes méthodes de Dynamique des Fluides Numériques: les solveurs de flux arbitrairement lagrangiens-eulériens (ALE) avec schémas explicites et implicites sont présentés et couplés au mouvement de maillage, l’adaptation de maillage feature-based instationnaire pour les géométries mobiles prend en compte les changements des connectivités de maillage durant toute la simulation, l'état adjoint est étendu aux problèmes de géométries mobiles et l'adaptation de maillage instationnaire orientée fonctionnelle pour les maillages mobiles est déduite d'une estimation d'erreur a priori. Plusieurs exemples numériques issus du secteur aéronautique et du domaine de sécurité civile sont considérés. / When dealing with CFD problems, mesh adaptation is interesting for its ability to approach the asymptotic convergence and to obtain an accurate prediction for complex flows at a lower cost. Anisotropic mesh adaptation method reduces the number of degrees of freedom required to reach a given solution accuracy, thus impact favorably the CPU time. Moreover, it reduces the numerical scheme dissipation by automatically taking into account the anisotropy of the physical phenomena inside the mesh. Two main approaches exist in the literature. Feature-based mesh adaptation which is mainly deduced from an interpolation error estimate using the Hessian of the chosen sensor controls the interpolation error of the sensor over the whole computational domain. Such approach is easy to set-up and has a wide range of application, but it does not take into account the considered PDE used to solve the problem. On the other hand, goal-oriented mesh adaptation, which focuses on a scalar output function, takes into consideration both the solution and the PDE in the error estimation thanks to the adjoint state. But, the design of such error estimate is much more complicated. This thesis presents the results obtained with different CFD methods : the Arbitrary Lagrangian Eulerian (ALE) flow solvers with explicit and implicit schemes are presented and coupled to the moving mesh process, the feature-based unsteady mesh adaptation for moving geometries takes into account the changes of connectivites during the whole simulation, the adjoint state is extended to moving geometries problems and goal-oriented unsteady mesh adaptation for moving meshes is derived from an a priori error estimate. Several numerical examples are considered in the aeronautics sector and the field of civil security.

Les simulations dépendant du temps sont toujours un challenge dans l'industrie, notamment à cause des problèmes posés par les géométries mobiles en termes de CPU et de précision. Cette thèse présente des contributions à certains aspects des simulations en géométrie mobile. Un algorithme de bouger de maillage fondé sur une déformation de maillage sur un grand pas de temps et des changements de connectivité (swaps) est étudié. Une méthode d'élasticité et une méthode d'interpolation directe sont comparées en 3D, démontrant l'efficacité de l'algorithme. Cet algorithme est couplé à un solver ALE, dont les schémas et l'implémentation en 3D sont décrits en détail. Une interpolation linéaire est utilisée pou traiter les swaps. Des cas de validation montrent que les swaps n'influent pas notablement sur la précision de la solution. Plusieurs examples complexes en 3D démontrent la puissance de cette approche, pour des mouvement imposés ou pour des problèmes d'Interaction Fluide-Structure. L'adaptation de maillage anisotrope a démontré son efficacité pour améliorer la précision des calculs stationnaires pour un coût raisonnable. On considère l'extension de ces méthodes aux problèmes instationnaires, en mettant à jour l'algorithme de point fixe précédent grâce à une ananlyse de l'erreur espace-temps fondée sur le modèle de maillage continu. Une parallélisation efficace permet de réaliser des simulations adaptatives instationnaires avec une précision inégalée. Cet algorithme est étendu au cas des géométries mobiles en corrigeant la métrique optimale instationnaire. Finalement, plusieurs exemples 3D de simulations adaptatives en géométries mobiles démontrent l'efficacité de l'approche. / Time dependent simulations are still a challenge for industry, notably due to problems raised by moving boundaries, both in terms of CPU cost and accuracy. This thesis presents contributions to several aspects of simulations with moving meshes. A moving-mesh algorithm based on a large deformation time step and connectivity changes (swaps) is studied. An elasticity method and an Inverse Distance Weighted interpolation method are compared on many 3D examples, demonstrating the efficiency of the algorithm in handling large geometry displacement without remeshing. This algorithm is coupled with an Arbitrary-Lagrangian-Eulerian (ALE) solver, whose schemes and implementation in 3D are described in details. A linear interpolation scheme is used to handle swaps. Validation test cases showed that the use of swaps does not impact notably the accuracy of the solution, while several other complex 3D examples demonstrate the capabilities of the approach both with imposed motion and Fluid-Structure Interaction problems. Metric-based mesh adaptation has proved its efficiency in improving the accuracy of steady simulation at a reasonable cost. We consider the extension of these methods to unsteady problems, updating the previous fixed-point algorithm thanks to a new space-time error analysis based on the continuous mesh model. An efficient p-thread parallelization enables running 3D unsteady adaptative simulations with a new level of accuracy. This algorithm is extended to moving mesh problems, notably by correcting the optimal unsteady metric. Finally several 3D examples of adaptative moving mesh simulations are exhibited, that prove our concept by improving notably the accuracy of the solution for a reasonable time cost.

This diploma thesis is dedicated to one of the problems in wireless sensor networks – over the air upgrade (OTAU) of nodes. The requirements for upgrade and possible ways how to transfer new firmware image to nodes are stated in the first chapter. The second chapter is focused on solving the problem of firmware dissemination to whole network based on analysis of known protocols. The practical part of this thesis deals with OTAU design and implementation in Lightweight Mesh software stack from Atmel. Proposed system was tested on deRFnode platform with plugged in deRFmega128 module. The upgrade process of designed system is evaluated on test network in last part. Practical test results are compared with OTAU solution for BitCloud.

Some boundary value problems yield anisotropic solutions, e.g. solutions with boundary layers. If such problems are to be solved with the finite element method (FEM), anisotropically refined meshes can be advantageous. In order to construct these meshes or to control the error one aims at reliable error estimators. For isotropic meshes such estimators are known but they fail when applied to anisotropic meshes. Rectangular (or cuboidal) anisotropic meshes were already investigated. In this paper an error estimator is presented for tetrahedral or triangular meshes which offer a much greater geometrical flexibility.