Control Techniques for Optimal Management of Microgrids in the Presence of Uncertainty

Recent years have witnessed an increasing adoption of renewables and electric vehicles (EVs) to overcome issues caused by carbon-fossil resources. In fact, renewable generation makes possible to produce clean energy without impact on the environment. On the other hand, EVs are capable of reducing gas emissions by fully exploiting these new resources. However, both of these technologies are an intrinsic source of uncertainty. Indeed, renewables represent uncontrollable and intermittent energy resources whose production strictly depends on meteorological conditions. Concerning EVs, their energy demand depends on several factors such as traffic conditions and user preferences. In this context, the concept of microgrid plays a key role, since it represents the simplest aggregation level of different components and players of the grid for optimal management and control of the electricity system in the presence of these uncertainties. Indeed, the development of optimization algorithms and novel power system protocols leads to a proper integration of EVs and renewables. In this setting, the aim of this thesis is to design novel control techniques for dealing with the uncertainty in microgrids. Two main aspects are considered. The first one is focused on providing solutions to overcome the uncertainty affecting EVs. The addressed problems are focused on reducing the daily peak power consumption, providing a competitive selling price and ensuring grid technical constraints. The second one is related to the optimal energy management of a smart building under environmental forecast uncertainty. In particular, the problem of cost-optimal operation of a centralized heating and air conditioning system is studied. Building participation in a Demand-Response program is also considered. Numerical simulations are provided to assess the performance and computational feasibility of the proposed solutions.