Energy Trading in Grid-connected PV-Battery Electric Vehicle Charging Station

As electric vehicles proliferate the cities and more charging stations get connected to the grid, the load demand on the grid increases and the available power systems infrastructure may not suffice to supply the demand. Grid-tied renewable energy sources-based charging stations have been proposed as a feasible solution to this problem. However, such an architecture comes with multiple stochastic variables to be considered including the charging station load, a dynamic grid tariff and the renewable energy generator output, that complicate optimal energy scheduling. Therefore, the development of scheduling algorithms that can adapt to such stochasticity is required. In this paper, the energy scheduling problem in a grid-connected PV-battery electric vehicle charging station is modeled as a Markov Decision Process and a Q-learning algorithm is used to solve it. The simulation results show that the algorithm successfully obtains a day-ahead energy schedule that reduces the cost of energy purchased from the grid while maximizing the revenue from energy sales to the grid.