Robust Operation of a Reconfigurable Electrical Distribution System by Optimal Charging Management of Plug-In Electric Vehicles Considering the Technical, Social, and Geographical Aspects

Abstract

This chapter proposes a robust approach to study the optimal operation problem of a reconfigurable electrical distribution system while optimally managing the charging/discharging patterns of plug-in electric vehicle (PEV) fleet considering their technical, social, and geographical aspects. Herein, it is assumed that the electrical system is highly penetrated by the renewable energy sources (RESs), and the total daily energy generated by the RESs is adequate for the daily electricity demand of system; however, an effective approach is necessary to reliably and economically operate it. The electrical distribution network includes the electrical loads, RESs, energy storage systems (ESSs), switches installed on the electrical feeders, and PEVs with the capabilities of vehicle-to-grid (V2G) and grid-to-vehicle (G2V). In this study, the drivers are grouped in three different social classes based on their income level, that is, low-income, moderate-income, and high-income. The behavior of each social class of drivers is modelled based on the social and geographical aspects including the drivers’ distance from a charging station (CHS) and the value of incentive to provide the V2G and G2V services at the suggested CHS and recommended period. The proposed approach includes the stochastic model predictive control (MPC) that stochastically, adaptively, and dynamically solves the problem and handles the variability and uncertainties concerned with the probabilistic power of RESs and drivers’ behavior. The simulation results demonstrate that applying the proposed approach can remarkably decrease the minimum operation cost of problem and enhance the system reliability. It is shown that the behavior of drivers can affect the optimal configuration of system, optimal status of ESSs, and even optimal scheme of PEV fleet management (FM). It is proven that the application of proposed approach guarantees the robustness of problem outputs with respect to the prediction errors.