Presently power systems are being operated under high stress level conditions unforeseen at the moment they were designed. These operating conditions have negatively impacted reliability, controllability and security margins. FACTS devices and HVDC transmissions have emerged as solutions to help power systems to increase the stability margins. VSC-HVDC transmissions are of particular interest since the principal characteristic of this type of transmission is its ability to independently control active power and reactive power. This thesis presents various control strategies to improve damping of electromechanical oscillations, and also enhance transient and voltage stability by using VSC-HVDC transmissions. These control strategies are based of different theory frames, namely, modal analysis, nonlinear control (Lyapunov theory) and model predictive control. In the derivation of the control strategies two models of VSC-HVDC transmissions were also derived. They are Injection Model and Simple Model. Simulations done in the HVDC Light Open Model showed the validity of the derived models of VSC-HVDC transmissions and the effectiveness of the control strategies. Furthermore the thesis presents an analysis of local and remote information used as inputs signals in the control strategies. It also describes an approach to relate modal analysis and the SIME method. This approach allowed the application of SIME method with a reduced number of generators, which were selected based on modal analysis. As a general conclusion it was shown that VSC-HVDC transmissions with an appropriate input signal and control strategy was an effective means to improve the system stability. / QC 20110412

Les travaux de cette thèse portent sur la stabilisation des systèmes affines commutés. Ces systèmes appartiennent à la classe des systèmes dynamiques hybrides. Ils possèdent de plus la particularité d'avoir des points de fonctionnement non auto-maintenables : il n'existe pas de loi de commutations permettant de maintenir l'état du système en ce point. De ce fait, la stabilisation de ces systèmes en imposant à la loi de commutations une durée minimale entre chaque commutation aboutit à une convergence des trajectoires dans une région de l'espace d'état. Après avoir synthétisé différentes stratégies de commutations échantillonnées construites à partir d'une fonction de commande de Lyapunov en temps continu, nous cherchons à déterminer la région de l'espace dans laquelle converge asymptotiquement l'ensemble des trajectoires du système. Par la résolution d'un problème d'optimisation, une estimation de la taille de cette région est donnée et un lien avec les incertitudes du système y est établi. Un second problème de stabilisation est étudié dans cette thèse, en considérant une stratégie de commande basée observateur par retour de sortie. Cependant, du fait de la nature hybride du système, son observabilité est directement liée à la séquence de commutations. Il est alors nécessaire de garantir à la fois l'observabilité, par une condition algébrique, et la convergence du système vers un point de fonctionnement, par l'existence d'une fonction de commande de Lyapunov / This PhD thesis deals with the stabilization of switched affine systems. These systems belong to the class of hybrid dynamical systems. They exhibit a particular behavior: no switching law exists such that the state can be maintained on a chosen operating point. Hence, assuming a dwell time condition on switchings exists, the stabilization of these systems leads to a convergence of the trajectories to a region of the state space. Based on a control Lyapunov function in continuous time, we synthesize several sampled-data switching strategies. The whole trajectories asymptotically converge to a region which we attempt to determine. Solving an optimization problem, an estimation of the size of this region is given. A link with the system uncertainties is also established. This PhD thesis is dedicated to a second stabilization issue: observer-based output-feedback synthesis. By its hybrid nature, the observability of the system is connected to the switching sequence. Therefore, the synthesis of the switching strategy must respect an observability condition and guarantee the convergence to the operating point. The observability is achieved thanks to an algebraic condition. The convergence property is based on the existence of a control Lyapunov function

Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro / Neste trabalho apresentamos as etapas para a utilização do método da Programação Dinâmica, ou Princípio de Otimização de Bellman, para aplicações de controle ótimo. Investigamos a noção de funções de controle de Lyapunov (FCL) e sua relação com a estabilidade de sistemas autônomos com controle. Uma função de controle de Lyapunov deverá satisfazer a equação de Hamilton-Jacobi-Bellman (H-J-B). Usando esse fato, se uma função de controle de Lyapunov é conhecida, será então possível determinar a lei de realimentação ótima; isto é, a lei de controle que torna o sistema globalmente assintóticamente controlável a um estado de equilíbrio. Como aplicação, apresentamos uma modelagem matemática adequada a um problema de controle ótimo de certos sistemas biológicos. Este trabalho conta também com um breve histórico sobre o desenvolvimento da Teoria de Controle de forma a ilustrar a importância, o progresso e a aplicação das técnicas de controle em diferentes áreas ao longo do tempo. / This dissertation presents the steps for using the method of Dynamic Programming or Bellman Optimization Principle for optimal control applications. We investigate the notion of control-Lyapunov functions (CLF) and its relation to the stability of autonomous systems with control. A control-Lyapunov function must satisfy the Hamilton-Jacobi- Bellman equation (H-J-B). Using this fact, if a control-Lyapunov function is known, it is possible to determine the optimal feedback law, in other words, the control law which makes the system globally asymptotically controllable at an equilibrium state. As an application, we present a mathematical model suitable for an optimal control problem of certain biological systems. This dissertation also presents a brief historic about the development of the Control Theory in a way of illustrate the importance and the progress of the control techniques, specially where it can be applied, according to the diverse areas and different times that this techniques were discovered and used.

Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro / Neste trabalho apresentamos as etapas para a utilização do método da Programação Dinâmica, ou Princípio de Otimização de Bellman, para aplicações de controle ótimo. Investigamos a noção de funções de controle de Lyapunov (FCL) e sua relação com a estabilidade de sistemas autônomos com controle. Uma função de controle de Lyapunov deverá satisfazer a equação de Hamilton-Jacobi-Bellman (H-J-B). Usando esse fato, se uma função de controle de Lyapunov é conhecida, será então possível determinar a lei de realimentação ótima; isto é, a lei de controle que torna o sistema globalmente assintóticamente controlável a um estado de equilíbrio. Como aplicação, apresentamos uma modelagem matemática adequada a um problema de controle ótimo de certos sistemas biológicos. Este trabalho conta também com um breve histórico sobre o desenvolvimento da Teoria de Controle de forma a ilustrar a importância, o progresso e a aplicação das técnicas de controle em diferentes áreas ao longo do tempo. / This dissertation presents the steps for using the method of Dynamic Programming or Bellman Optimization Principle for optimal control applications. We investigate the notion of control-Lyapunov functions (CLF) and its relation to the stability of autonomous systems with control. A control-Lyapunov function must satisfy the Hamilton-Jacobi- Bellman equation (H-J-B). Using this fact, if a control-Lyapunov function is known, it is possible to determine the optimal feedback law, in other words, the control law which makes the system globally asymptotically controllable at an equilibrium state. As an application, we present a mathematical model suitable for an optimal control problem of certain biological systems. This dissertation also presents a brief historic about the development of the Control Theory in a way of illustrate the importance and the progress of the control techniques, specially where it can be applied, according to the diverse areas and different times that this techniques were discovered and used.