Stochastic Model Predictive Control Framework For Electric Vehicles Charging In An Overloaded Distribution Network

Abstract

In this thesis work a control problem is formulated to handle the charging of Electric Vehicles (EVs) in a power distribution system which is prone to network overloading. Such overloading results due to the lack of capacity of the distribution system to deliver power to its consumers, and it is prevalent in the developing countries, where the addition of EVs load is expected to further aggravate it. Furthermore, the addition of EVs in a distribution network introduces multiple uncertainties, e.g. uncertainties in the arrival times of the EVs to Charging Stations (CSs), their initial battery State Of Charges (SOCs) and their charging time requirements, which adds to the load fluctuation and results in the instability of the grid. To address these issues, we formulate the problem of charging EVs in a stochastic framework where a Stochastic Model Predictive Control (S-MPC) based approach is proposed. The approach is designed to simultaneously achieve the desired SOCs of the batteries of the EVs, standing at the CSs, and minimize network overloading and load fluctuation in the grid. The simulated results show that the approach successfully achieves the desired objectives in the presence of the said uncertainties.