Online energy management of PV-assisted charging station under time-of-use pricing

• Energy management leveraging both offline optimization and online decision-making.
• The method is deployed without forecasting on the PV power and charging demand.
• Operational profit, solar consumption, and charging completeness are considered.
• The method provides some insight into the development of EMS on the demand side.

In order to maximize the operation profit while maintaining the service quality, an effective energy management scheme is highly needed for the Photovoltaic-assisted Charging Station (PVCS). Considering the uncertainty of Electric Vehicle (EV) charging demand and PV power output, it will be challenging to determine the charging power for EVs to make informed real-time decisions. In this study, an online energy management method leveraging both offline optimization and online learning is proposed. In order to maximize the self-consumption of Photovoltaic (PV) energy and decide the power supplied from the power grid with Time-of-Use (TOU) pricing, here online learning is coupled with the rule-based decision-making to obtain a real-time online algorithm. The knowledge base for online learning is derived and updated from the results of offline optimization after every operation day. The PVCS located at workplace parking lots is used as an example to test the proposed method. The simulation results show that the method can be implemented without the information on future PV power and charging demand. The obtained results are close to the optimal results from offline optimization.