EV charging load scheduling with high wind power penetration: A robust shortest path approach

Load scheduling is an effective way to improve the wind power penetration. Electric vehicle (EV) is an important type of the controllable load for its elasticity of charging time. However, the charging load of the EVs is random by nature and the accurate probability distributions of the wind power and EV parking event are hard to achieve. Therefore, in this paper, we consider scheduling EV charging load to improve the wind power penetration and reduce the charging cost. We make the following contributions. First, the EV load scheduling problem is formulated as the robust shortest path Markov Decision Process (MDP) to obtain a robust charging policy. The charging policy can reduce the charging cost even under the worst-case scenario of the wind power and EV parking event. Second, a Simulation-based Policy Improvement (SBPI) algorithm is developed to solve this robust shortest path problem. The robustness and the reduced charging cost are demonstrated through numerical results.