Game-theoretic Control of Autonomous Power Grids

Power systems are currently undergoing a period of unprecedented transformations. Environmental and sustainability concerns lead to the replacement of centralized generation, based on conventional fossil fuel-based power plants, with distributed generation from renewable energy sources. In addition, a variety of new autonomous entities able to adjust their load demand or provide ancillary services to the grid are increasing the complexity of energy systems, requiring a decentralization of the control structures. In this context, autonomous power grids have been proposed as a necessary paradigm to capture the need for distributed operation of power grids. Nevertheless, the existing control and optimization techniques are inadequate to reach this goal while ensuring the efficiency and security of power systems. Due to its capacity to capture interactions among interdependent decision-making entities, game theory offers a promising way to implement and control autonomous power grids. Nonetheless, several technical issues must be solved in order to fully implement this new paradigm. As a result, this thesis is dedicated to solving two of the most important research challenges in designing and operating game-theoretical control for autonomous power grids. In the first part, this thesis deals with the development of optimization tools devoted to closing the gap between variable generation and adjustable load demand. The increasing penetration of renewable energy sources poses significant challenges to the existing power systems due to the difficulty in coping with their inherent time-varying nature. To this aim, a series of stochastic techniques for energy system optimization are proposed to accommodate uncertainty in autonomous power grids operation. Consequently, game-theoretical frameworks are defined with the aim of increasing the flexibility of the grid through the active participation of autonomous entities. The second part of this thesis is further focused on coordinating autonomous entities in game-theoretical frameworks. The coordination of these entities is not straightforward due to the fact that they are interconnected through power lines and therefore must respect the so-called power flow constraints. The nonconvexity of the resulting control problem thus makes difficult to use traditional mathematical tools. To solve this problem, a novel mathematical theory for game-theoretical frameworks with nonconvex coupling constraints is developed and applied to ensure the quality and feasibility of autonomous power grid operation.