Planning the operation and expansion of power distribution systems considering electric vehicles (smart charging)

Abstract

In conventional power systems, the main role of distribution networks has been to pass the electricity from transmission lines to small-size consumers. In that structure, distribution networks function as passive intermediator between the transmission sector and small electric loads. Notwithstanding, due to the various advantages distributed generation (DG) units can offer to power systems, including enhancement in reliability, sustainability, and resiliency of distribution networks, this structure has been revolutionized during the last few years. Nowadays, distribution networks accommodate numerous DG units and have enough capability to actively interact with the transmission sector to offer ancillary and flexibility services and contribute to cost reduction in electricity generation and the upgrade of transmission and distribution networks. Moreover, the proliferation of photovoltaic and wind turbine units and the emergence of plug-in electric vehicles have complicated the planning of distribution networks since these technologies pose high-level uncertainties linked to renewable energy generation and the future growth of electricity demand. As a result, a deterministic mixed-integer linear programming (MILP) model is developed in this chapter to simultaneously embrace the expansion planning of distribution networks and the allocation of electric vehicle charging stations. Then, the aforementioned deterministic MILP optimization model is transformed into a stochastic formulation to include uncertainties linked to renewable energy generation and electric loads. The respective results are compared and discussed through some case studies indicating that charging stations are located adjacent to DG units. Furthermore, it can be concluded that the stochastic model could help to effectively address uncertainties, providing reliable solutions for the optimal multistage planning of distribution networks, although slightly increasing the total cost compared to the deterministic approach.