Deep Q-network based battery energy storage system control strategy with charging/discharging times considered

作者：Jun Cai, Maowen Fu, Ying Yan, Zhong Chen, Xin Zhang · 发表于：Applied Energy · 年份：2025 · DOI：10.1016/j.apenergy.2025.126384 · 被引用次数：7 · 研究领域：Microgrid Control and Optimization、Advanced Battery Technologies Research、Smart Grid Energy Management

The Battery Energy Storage System (BESS) plays a pivotal role in maintaining the balance of electricity supply and demand on the user side. This paper proposes an energy management system (EMS) for the BESS based on the Deep Q-Network (DQN) algorithm that takes into account the battery charging and discharging times. Initially, a mathematical model of the EMS is established. Subsequently, the optimal decision-making process of EMS is formulated as Markov Decision Process (MDP), and based on this, the MDP formula and DQN algorithm are designed to design charging/discharging schedules based on load conditions. Finally, an experimental study was conducted based on the actual load data of a certain line in Zunyi, Guizhou, China. The test results show that the optimization method proposed in this study reduces the maximum variance of power grid fluctuations to 49 % of the original variance, while reducing the number of battery charging and discharging cycles to the range of 1/3 to 1/2 of the initial value. This delays the battery aging process, improving the economic and practical efficiency of energy management strategies.