Smart charge-optimizer: Intelligent electric vehicle charging and discharging

Abstract

The important steps toward a low-carbon economy and sustainable energy future is switch to Electric Vehicles(EVs).The rapid development of EVs has brought a risk to reliability of the electrical system. However, the high electricity consumption of EVs will lead to the overload of power grid transformers. Strategies for scheduling charging and discharging that work are essential to reducing the negative grid effects of EVs. In order to reduce the overload of power grid transformers, this paper explores two strategies for intelligent charging and discharging scheduling. The first one is Long Short-Term Memory coupled with Integer Linear Programming(LSTM-ILP)and the second one is Q-learning. The LSTM-ILP aims to minimize the charging and discharging schedules delay. The Q-learning method makes use of reinforcement learning to ascertain the best course of action for EVs in relation to their state-of-charge and the demand on the grid. The outcomes of this research show that both strategies are successful in lowering the peak-to-average ratio of the grid and lessening the influence of EV charging demands. • This research aims to Couple Long Short-Term Memory with Integer Linear Programming • Applying Q-learning to minimize the peak to-average ratio of grid load through effective peak shaving and valley filling • Minimizing EV charging costs for users while respecting their mobility needs

Keywords

• 1. Long Short-Term Memory with Integer Linear Programming (LSTM-ILP)
• 2. Q-learning