[pt] Em 2015 quase metade da população mundial vivia em área de risco de transmissão de malária. Neste mesmo ano, estimam-se 214 milhões de casos e 438 mil fatalidades. A principal forma de prevenção e redução da transmissão da malária é através do controle dos vetores, em particular, destaca-se o uso de mosquiteiros impregnados com inseticidas de longa duração (MILD). Neste contexto, os programas de distribuição de MILDS enfrentam desafios relacionados a obtenção de fundos e à gestão da cadeia de suprimentos como, por exemplo, incertezas associadas as atividades logísticas, as variáveis de oferta e demanda, e a volatilidade de preços. À luz destes fatos, esta dissertação propõe um modelo de otimização robusta, fundamentado em extensões dos arcabouços teóricos de Bertsimas e Sim (2004) e Fernandes et al. (2016), capaz de minimizar os custos de um programa de distribuição de mosquiteiros ou, dada uma restrição orçamentária, maximizar a distribuição para áreas prioritárias. Ademais, foi realizada uma revisão da literatura acadêmica acerca de modelos de otimização robusta aplicados no contexto da logística humanitária, onde alguns aspectos ainda pouco explorados foram ressaltados e considerados no modelo proposto. Um estudo de caso real é feito sobre um projeto feito do Fundo das Nações Unidas para crianças na Costa do Marfim. Os resultados apontam que conforme esperado, à medida que o nível de robustez considerado no modelo cresce, os custos totais também aumentam. Em contrapartida, o modelo robusto fornece soluções com maior flexibilidade na cadeia de suprimentos para a eventual necessidade de se ajustar os planos de compras e distribuição. Por fim, as soluções robustas foram avaliadas através de simulações de Monte Carlo, indicando que, conforme desejado, a probabilidade de viabilidade dos planos aumentam junto com nível de conservadorismo da solução. / [en] In 2015, almost half of the world population lived in areas at risk of malaria transmission. There were around 214 million malaria cases and 438,000 associated deaths. One of the major paths to prevent and reduce malaria transmission is through vector control, especially with the use of insecticide-treated nets (ITN). In this context, ITN distribution campaigns face several challenges, such as uncertainties related to funding, transportation, market and price volatility, which might be effectively tackled through long-term agreements and proper planning. However, that might not be an option for all humanitarian organizations and governments. Besides, considering uncertainties during budgetary planning is particular relevant. In this sense, a robust optimization model, based on Bertsimas and Sim (2004) and Fernandes et al. (2016) frameworks, is proposed to minimize the involved costs or, given a budget constraint, maximize the coverage of priority areas. A literature review on robust optimization applied to humanitarian logistics is conducted, in which aspects with less academic research attention are revealed and considered in the model, such as the simultaneous account of the aforementioned uncertainties and demand prioritization. A United Nations Children s Fund campaign in Ivory Coast is studied, and reveals that, as expected, as the robustness level increases so does the total costs. In return, the robust model generally provides a solution with improved supply chain flexibility, that might minimize efforts, in case it is necessary to adjust procurement and transportation plans when uncertainty is revealed. In addition, robust solutions were assessed through Monte Carlo simulations against several realizations of uncertain parameters values, pointing that, as desired, solution feasibility increases alongside the specified level of conservatism.

abstract: Corrective transmission topology control schemes are an essential part of grid operations and are used to improve the reliability of the grid as well as the operational efficiency. However, topology control schemes are frequently established based on the operator's past knowledge of the system as well as other ad-hoc methods. This research presents robust corrective topology control, which is a transmission switching methodology used for system reliability as well as to facilitate renewable integration. This research presents three topology control (corrective transmission switching) methodologies along with the detailed formulation of robust corrective switching. The robust model can be solved off-line to suggest switching actions that can be used in a dynamic security assessment tool in real-time. The proposed robust topology control algorithm can also generate multiple corrective switching actions for a particular contingency. The solution obtained from the robust topology control algorithm is guaranteed to be feasible for the entire uncertainty set, i.e., a range of system operating states. Furthermore, this research extends the benefits of robust corrective topology control to renewable resource integration. In recent years, the penetration of renewable resources in electrical power systems has increased. These renewable resources add more complexities to power system operations, due to their intermittent nature. This research presents robust corrective topology control as a congestion management tool to manage power flows and the associated renewable uncertainty. The proposed day-ahead method determines the maximum uncertainty in renewable resources in terms of do-not-exceed limits combined with corrective topology control. The results obtained from the topology control algorithm are tested for system stability and AC feasibility. The scalability of do-not-exceed limits problem, from a smaller test case to a realistic test case, is also addressed in this research. The do-not-exceed limit problem is simplified by proposing a zonal do-not-exceed limit formulation over a detailed nodal do-not-exceed limit formulation. The simulation results show that the zonal approach is capable of addressing scalability of the do-not-exceed limit problem for a realistic test case. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2015

abstract: In recent years, wide bandgap (WBG) devices enable power converters with higher power density and higher efficiency. On the other hand, smart grid technologies are getting mature due to new battery technology and computer technology. In the near future, the two technologies will form the next generation of smart grid enabled by WBG devices. This dissertation deals with two applications: silicon carbide (SiC) device used for medium voltage level interface (7.2 kV to 240 V) and gallium nitride (GaN) device used for low voltage level interface (240 V/120 V). A 20 kW solid state transformer (SST) is designed with 6 kHz switching frequency SiC rectifier. Then three robust control design methods are proposed for each of its smart grid operation modes. In grid connected mode, a new LCL filter design method is proposed considering grid voltage THD, grid current THD and current regulation loop robust stability with respect to the grid impedance change. In grid islanded mode, µ synthesis method combined with variable structure control is used to design a robust controller for grid voltage regulation. For grid emergency mode, multivariable controller designed using H infinity synthesis method is proposed for accurate power sharing. Controller-hardware-in-the-loop (CHIL) testbed considering 7-SST system is setup with Real Time Digital Simulator (RTDS). The real TMS320F28335 DSP and Spartan 6 FPGA control board is used to interface a switching model SST in RTDS. And the proposed control methods are tested. For low voltage level application, a 3.3 kW smart grid hardware is built with 3 GaN inverters. The inverters are designed with the GaN device characterized using the proposed multi-function double pulse tester. The inverter is controlled by onboard TMS320F28379D dual core DSP with 200 kHz sampling frequency. Each inverter is tested to process 2.2 kW power with overall efficiency of 96.5 % at room temperature. The smart grid monitor system and fault interrupt devices (FID) based on Arduino Mega2560 are built and tested. The smart grid cooperates with GaN inverters through CAN bus communication. At last, the three GaN inverters smart grid achieved the function of grid connected to islanded mode smooth transition / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2017

L'objectif principale de cette thèse est de développer contrôleurs innovants MIMO pour la dynamique véhicule tout en préservant la stabilité du véhicule dans les situations de conduite critiques. Des stratégies de commandes innovatrices ont été introduites pour résoudre cette problématiques. En effet, ces travaux se base sur travaux l'utilisation de la commande LPV/Hinf pour contrôler simultanément les actionneurs de freinage, braquage et de suspensions pour réaliser les objectives du contrôle.Aussi de stratégies d'estimation du profil de route très intéressant et qui peuvent apporter une solution industrielle très intéressante pour développer des contrôleurs qui assurent adaptative aux différentes conditions de route.Aussi des stratégies de commande tolérante aux défauts actionneurs ont été établi en exploitant les caractéristiques de la commande LPV pour compenser la perte de certains actionneurs (en sachant que la voiture est un système sur actionné). Des implémentations ont été effectuées sur des bancs de test et sur un véhicule réel pour prouver l'efficacité des stratégies. / The main issue of this thesis is to work out new Global Chassis MIMO controllers that enhance the overall dynamics of the vehicle while preserving the vehicle stability in critical driving situations. Many innovative strategies have been explored and finalized to deal with these problematics. Various solutions have been given to deal with the vehicle stability and performance objectives. Indeed, many works based on the LPV/Hinf approach have been developed to control simultaneously the braking, steering and suspension actuators. On the other hand, innovative road profile estimation strategies have been introduced and validated via experimental procedures, providing new cheap and easily implementable techniques to estimate the road profile characteristics. Then, the vehicle control is adapted, depending on the road roughness (since it influences greatly the behaviour and the stability of the car). Several fault tolerant control strategies have been also considered to handle the actuators failures while keeping the vehicle stability, safety and enhancing the dynamical behaviour of the car in dangerous and critical driving situations.The general content of this thesisis as follows :-PART I : Theoretical backgrounds and vehicle modeling.-PART II : Road adaptive control vehicle dynamics.-PART III : Global chassis control using several actuators.Also, during this thesis and using the previous works of the advisors and the thesis results, a Matlab ToolBox "Automotive" has been developed to provide a bench test for the different automotive control studies. Implementations on test beds and real vehicle are also achieved to prove the efficiency of the proposed strategies.

Les travaux présentés dans cette thèse portent sur la conception robuste de produit et plus particulièrement sur la phase de pré-dimensionnement dans le cas où un modèle de dimensionnement et un cahier des charges sont déjà définis. Une approche pour réaliser de l’optimisation robuste est proposée pour réduire la dispersion de la fonction objectif du cahier des charges du produit lorsque les paramètres de conception sont sujets aux incertitudes, conserver une bonne performance du produit et assurer une faisabilité des contraintes. Nous proposons ainsi la formulation d’un cahier des charges dit « robuste » transformant la fonction objectif et les contraintes du cahier des charges initial afin d’intégrer une notion de robustesse préalablement définie. La seconde contribution est une analyse des méthodes trouvées dans la littérature pour la propagation d’incertitudes à travers des modèles de dimensionnement. Les variations des paramètres sont alors modélisées par des dispersions probabilistes. L’analyse théorique du fonctionnement de chaque méthode est complétée par des tests permettant d’étudier la précision des résultats obtenus et de sélectionner la méthode utilisée par la suite. L’approche pour l’optimisation robuste de produit proposée dans ce travail est finalement mise en œuvre et testée sur deux études de cas. Elle intègre la méthode de propagation d’incertitudes dans une boucle de l’algorithme d’optimisation de manière à automatiser la recherche d’une solution optimale robuste pour le dimensionnement du produit. / The work presented in this thesis deals with the robust design of products. Particularly, it focuses on the design process preliminary phase where design models and specifications are already defined. A robust optimization approach is proposed. It aims to: reduce the scattering of the objective function included in the product specifications when the uncertainties reach the design parameters, maintain good performance of the product and ensure the constraints feasibility. We propose a new “robust” product specification that changes the objective function and the constraints of the initial specification in order to integrate the concept of robustness previously defined. The second contribution is an analysis of methods found in literature to propagate uncertainties across design models. Design parameter variations are modeled by probability distributions. The theoretical analysis of these methods is completed by numerous tests to investigate the accuracy of the results and to select the method used thereafter. The robust product optimization approach proposed in this work is finally implemented and tested on two case studies. It incorporates the propagation of uncertainties within the optimization loop to automate the search of a robust optimal solution for the design product.

Orientador: Vicente Lopes Júnior / Banca: Walter Katsumi Sakamoto / Banca: Alberto Luiz Serpa / Resumo: Este trabalho apresenta o projeto do regulador linear quadrático (LQR - do inglês Linear Quadratic Regulator) para atenuar vibrações em estruturas mecânicas. Estas estruturas, com atuadores e sensores acoplados, são denominadas estruturas inteligentes. Os projetos de controladores ativos são resolvidos utilizando desigualdades matriciais lineares (LMIs - do inglês Linear Matrix Inequalities). Assim, é possível projetar controladores robustos considerando incertezas paramétricas na planta a ser controlada. São utilizados atuadores e sensores piezelétricos (PZTs) para aplicações em estruturas flexíveis dos tipos vigas e placas e, também, atuadores de pilha para aplicações em estruturas do tipo treliça. O problema do posicionamento ótimo dos atuadores e sensores piezelétricos também é resolvido utilizando as normas de sistemas H2, H , Hankel e as matrizes grammianas de observabilidade e controlabilidade. O modelo matemático da estrutura inteligente é obtido a partir do Método dos Elementos Finitos e, também, utilizando o Método de Identificação de Subespaços através de dados experimentais. O problema de posicionamento ótimo dos atuadores e sensores e o controle ativo de vibração são apresentados em simulações numéricas e experimentais. Os resultados mostram que os controladores robustos aumentam o amortecimento estrutural minimizando as amplitudes de vibração. / Abstract: This work presents the Linear Quadratic Regulator design to vibration attenuation in mechanical structures. These structures are named Smart Structures because they use actuators and sensors electromechanically coupled. Active controller designs are solved using Linear Matrix Inequalities. So, it is possible to consider polytopic uncertainties. Piezoelectric actuators and sensors are used for applications in flexible structures as beams and plates and, also, stack actuators for applications in truss structures. Optimal placement problem of piezoelectric actuators and sensors also solved using H2, H , Hankel system norms and controllability and observability grammian matrices. The mathematical model of the smart structure is obtained through Finite Element Method and, also, through Numerical State Space of Subspace System Identification (Subspace Method) by experimental data. The optimal placement of actuator and sensor and the active vibration control is numerically and experimentally implemented. Results show that the robust controllers increase the structural damping minimizing magnitude of vibrations. / Mestre

abstract: In the three phases of the engineering design process (conceptual design, embodiment design and detailed design), traditional reliability information is scarce. However, there are different sources of information that provide reliability inputs while designing a new product. This research considered these sources to be further analyzed: reliability information from similar existing products denominated as parents, elicited experts' opinions, initial testing and the customer voice for creating design requirements. These sources were integrated with three novels approaches to produce reliability insights in the engineering design process, all under the Design for Reliability (DFR) philosophy. Firstly, an enhanced parenting process to assess reliability was presented. Using reliability information from parents it was possible to create a failure structure (parent matrix) to be compared against the new product. Then, expert opinions were elicited to provide the effects of the new design changes (parent factor). Combining those two elements resulted in a reliability assessment in early design process. Extending this approach into the conceptual design phase, a methodology was created to obtain a graphical reliability insight of a new product's concept. The approach can be summarized by three sequential steps: functional analysis, cognitive maps and Bayesian networks. These tools integrated the available information, created a graphical representation of the concept and provided quantitative reliability assessments. Lastly, to optimize resources when product testing is viable (e.g., detailed design) a type of accelerated life testing was recommended: the accelerated degradation tests. The potential for robust design engineering for this type of test was exploited. Then, robust design was achieved by setting the design factors at some levels such that the impact of stress factor variation on the degradation rate can be minimized. Finally, to validate the proposed approaches and methods, different case studies were presented. / Dissertation/Thesis / Doctoral Dissertation Industrial Engineering 2014

La simulation de systèmes d'ingénierie non linéaire complexes tels que les écoulements de fluide compressibles peut être ciblée pour rendre plus efficace et précise l'approximation d'une quantité spécifique (scalaire) d'intérêt du système. En mettant de côté l'erreur de modélisation et l'incertitude paramétrique, on peut y parvenir en combinant des estimations d'erreurs axées sur des objectifs et des raffinements adaptatifs de maillage spatial anisotrope. A cette fin, un cadre élégant et efficace est celui de l'adaptation dite basé-métrique où une estimation d'erreur a priori est utilisée comme indicateur d’adaptation de maillage. Dans cette thèse on propose une nouvelle extension de cette approche au cas des approximations de système portant une composante stochastique. Dans ce cas, un problème d'optimisation est formulé et résolu pour un meilleur contrôle des sources d'erreurs. Ce problème est posé dans le cadre continu de l'espace de métrique riemannien. Des développements algorithmiques sont également proposés afin de déterminer les sources dominates d’erreur et effectuer l’adaptation dans les espaces physique ou des paramètres incertains. L’approche proposé est testée sur divers problèmes comprenant une entrée de scramjet supersonique soumise à des incertitudes paramétriques géométriques et opérationnelles. Il est démontré que cette approche est capable de bien capturé les singularités dans l’escape stochastique, tout en équilibrant le budget de calcul et les raffinements de maillage dans les deux espaces. / The simulation of complex nonlinear engineering systems such as compressible fluid flows may be targeted to make more efficient and accurate the approximation of a specific (scalar) quantity of interest of the system. Putting aside modeling error and parametric uncertainty, this may be achieved by combining goal-oriented error estimates and adaptive anisotropic spatial mesh refinements. To this end, an elegant and efficient framework is the one of (Riemannian) metric-based adaptation where a goal-based a priori error estimation is used as indicator for adaptivity. This thesis proposes a novel extension of this approach to the case of aforementioned system approximations bearing a stochastic component. In this case, an optimisation problem leading to the best control of the distinct sources of errors is formulated in the continuous framework of the Riemannian metric space. Algorithmic developments are also presented in order to quantify and adaptively adjust the error components in the deterministic and stochastic approximation spaces. The capability of the proposed method is tested on various problems including a supersonic scramjet inlet subject to geometrical and operational parametric uncertainties. It is demonstrated to accurately capture discontinuous features of stochastic compressible flows impacting pressure-related quantities of interest, while balancing computational budget and refinements in both spaces.

Conselho Nacional de Desenvolvimento CientÃfico e TecnolÃgico / Este trabalho visa a aplicaÃÃo dos controles robustos D-LQI e D-AlocaÃÃo de polos otimizados via LMIs em um conversor boost de alto ganho de tensÃo com cÃlula de comutaÃÃo de trÃs estados. Este conversor consiste numa topologia moderna derivada do conversor boost clÃssico. O boost à considerado um elevador de tensÃo, o qual converte uma entrada na faixa de 42-54V à 400 V. O conversor boost proposto à reduzido ao modelo de um conversor equivalente e à modelado no espaÃo de estados mÃdio, em que à observado que a matriz D diferente de zero, sendo entÃo uma modelagem que apresenta uma peculiaridade de acordo com a literatura, pois a soluÃÃo de controle à mais complexa. As estratÃgias de controle aplicadas usam de procedimentos matemÃticos denominados de Desigualdades Matriciais Lineares (LMIs-Linear Matrix Inequalities), que podem ser resolvidos por otimizaÃÃo convexa ou programaÃÃo semidefinida positiva (SDP procedures). As ferramentas matemÃticas utilizadas para resoluÃÃo das LMIs neste trabalho sÃo o Yalmip e Sedumi, que sÃo inseridas no MATLAB . AlÃm disso sÃo analisadas as incertezas presentes no processo, bem como a robustez do modelo em malha fechada. SÃo obtidos os resultados de simulaÃÃo via MATLAB -PSIM e sÃo feitas as anÃlises referentes a estes resultados, alÃm da anÃlise dos resultados experimentais e a conclusÃo do estudo, alÃm das propostas de trabalhos futuros. O ApÃndice mostra os procedimentos de instalaÃÃo dos resolvedores alÃm do uso correto com base nas equaÃÃes descritas na teoria sobre LMIs. / This work involves the application of robust controls D-LQI and D-pole placement via LMIs in a high-gain boost with three states switching cell. This converter consists of a modern topology derived the classic boost converter . This boost converter is considered a step-up converter, which a range of 42-54V voltage input to 400V voltage output. The proposed boost converter is reduced to equivalent model and is modeled at space state avarage, in which is observed that the matrix D not equal nought, being then a modeling that presents a peculiarity according to literature, thus the control solution is more complex. The control strategies applied use mathematical procedures called Linear Matrix Inequalities (LMIs), which can be solved by convex optimization or positive semidefinite procedures (SDP). The mathematical tools used to solve the LMIs this work are Yalmip and Sedumi, which are inserted in MATLAB. Further analyzes the uncertainties present in the process, as well as the robustness of closed loop model. The simulation results are obtained via MATLAB and PSIM and analyzes made regarding these results, besides the analysis of experimental results and conclusion of study, in addition to proposals for future work. The Appendix shows the installation procedures and use correct solvers based on the equations described in LMI theory.

Escavadeiras hidráulicas são máquinas versáteis, amplamente utilizadas na construção civil e mineração. Máquinas melhores, mais produtivas, eficientes e que oferecem segurança ao operador são uma demanda constante da indústria. Devido a estes fatores, o controle para a automação de uma escavadeira hidráulica tem sido investigado. Este estudo tem o seu foco voltado para o controle do manipulador do equipamento, que é considerado como um dos elementos fundamentais para o desenvolvimento de uma escavadeira automática. Para desenvolver um sistema de controle viável, primeiramente, foi realizado a modelagem matemática dos subsistemas mecânico e hidráulico do manipulador; posteriormente esses modelos foram acoplados para representar a interação dos subsistemas. Todos os modelos desenvolvidos foram comparados com modelos de referência, obtidos a partir de softwares comerciais dedicados a modelagem de sistema dinâmicos. Tendo sido verificado a capacidade de representação física dos modelos, a fase de projeto do controlador para o manipulador foi iniciada. Para que o controlador seja eficiente, este deve ter duas propriedades essenciais: robustez para lidar com as incertezas e distúrbios severos, e adaptabilidade para lidar com um ambiente de operação altamente dinâmico. A fim de projetar um controlador que considera a dinâmica de cada subsistema do manipulador, a técnica de controle em cascata foi adotada. Esta consiste em dividir o sistema global em subsistemas, de tal forma que seja possível projetar um controlador para cada subsistema. Devido à complexidade do modelo matemático, técnicas avançadas de controle linear e não linear foram combinadas no projeto dos controladores dos subsistemas. O controlador sintetizado foi testado através de simulação numérica, em ambiente MATLAB/Simulink®, na execução de um ciclo completo de trabalho pelo manipulador. Os resultados obtidos foram considerados satisfatórios, mesmo na presença de incertezas, distúrbios severos e de ruídos. Posteriormente, na comparação desses resultados com os de outros controladores, ficou claro que o melhor desempenho foi obtido com o controlador proposto. Isto indica a possível aplicabilidade de tal controlador para a automação deste tipo de equipamento. / Hydraulic excavators are versatile machines, widely used in civil construction and in mining. Better, more productive, and efficient machines that offer operator safety are a constant industry demand. Due to these factors, the control for the automation of a hydraulic excavator has been investigated. This study focuses on the control of the equipment\'s manipulator, which is considered as one of the fundamental elements for the development of an automatic excavator. To develop a viable control system, first, the mathematical modeling of the mechanical and hydraulic subsystems of the manip-ulator was carried out; later these models were coupled to represent the interaction between the subsystems. All the developed models were compared with reference models, obtained from a commercial software dedicated to dynamic system modeling. Having verified the physical representation capacity of the analytical models, the de-sign phase of the controller was started. For the controller to be efficient, it must have two essential properties: robustness to deal with severe uncertainties and disturb-ances, and adaptability to handle a highly dynamic operating environment. To design a controller that considers the dynamics of each subsystem of the manipulator, the cascade control technique was adopted. This consists of dividing the global system into subsystems, in such a way that it is possible to design a controller for each sub-system. Due to the complexity of the mathematical model, advanced linear and non-linear control techniques were combined in subsystem controllers design. The synthe-sized controller was tested by numerical simulation, in MATLAB/Simulink® environ-ment, in the execution of a complete work operation by the manipulator. The results obtained were considered satisfactory, even in the presence of uncertainties, severe disturbances and noise. Subsequently, in the comparison of these results with those of others controllers, it was clear that the best performance was obtained with the pro-posed controller. This indicates the possible applicability of such a controller to the automation of this type of equipment.