Both electric power systems and the transportation sector are essential constituents to modern life, enhancing social welfare, enabling economic prosperity and ultimately providing well-being to the people. However, to mitigate adverse climatological effects of emitting greenhouse gases, a rigorous de-carbonization of both industries has been set on the political agenda in many parts of the world. To this end, electrifying personal vehicles is believed to contribute to an affordable and reliable energy model that provides tolerable environmental impact. Representing an inherently flexible electricity demand, plug-in electric vehicles (PEVs) promise to facilitate the integration of variable renewable energy sources. Yet, how should the PEVs’ system usage be ideally coordinated for providing benefits to electric power systems in the presence of resource scarcity? The thesis develops a model of an aggregation agent as the interface to the wholesale electricity generators, which is envisaged to be in charge of procuring energy in electricity markets, exposed to uncertainty in prices, fleet availability and demand requirements. This aggregator could coordinate the PEV charging either with direct load control (DLC), i.e., sending power set points to the individual vehicles, or with indirect load control (ILC), i.e., by sending retail price signals. Contributing to the technical literature this thesis has on the one hand proposed a twostage stochastic linear program for the PEV aggregator’s day-ahead and balancing decisions with DLC over a large fleet of PEVs, while accounting for conditional value at risk in the objective function. On the other hand, it has put forward a formulation of ILC coordination as a bi-level optimization problem given by mathematical programming with equilibrium constraints, in which 1) the upper level decisions on retail tariffs and optimal bidding in electricity markets are subject to 2) the lower level client-side optimization of PEV charging schedules. These decisions may respect a potential discomfort that could arise when PEV users have to deviate from their preferred charging schedule. Set in an existing, real medium voltage distribution network with urban characteristics and spatial PEV mobility, network UoS tariffs for capacity have been applied to both DLC and ILC scheduling by a PEV aggregator.
Universidad Pontificia Comillas. Madrid (España)
15 October 2015