MPC-based approach for online coordination of EVs Considering EV usage uncertainty

Abstract

This paper studies electric vehicle (EV) charging coordination in the presence of uncertainty. Considering that EVs usage are of considerable uncertainties, first, a probabilistic mechanism to model driving and charging patterns of EV owners utilizing Markov chain Monte Carlo (MCMC) is presented. Then a model predictive control (MPC) approach named online MPC (OL-MPC) is proposed to adaptively and intelligently coordinate the EV charging process which is able to cover both centralized and decentralized infrastructures concurrently. Technically, discrete-time manner, re-optimization characteristic, looking ahead and signal tracking are the main features of the proposed method which make it well-suited to address high uncertainties concerning EV usage. The proposed method will improve the performance of the grid under study, in general, and reduce costs, in particular. In order to evaluate the efficiency of the proposed method, first a conceptual analysis for OL-MPC performance validation is conducted and then the impacts of different benchmark scenarios are investigated by simulating on a modified IEEE 31-bus test system.