摘要
为提高配电网日网损减少收益、日峰谷价差套利收益,需合理分配电池充放电功率,为此,提出基于深度强化学习的分布式储能电池能量调度方法。构建了分布式储能电池剩余可用能量估计模型,使用基于深度强化学习的电池能量调度模型,将目标函数设为日网损减少收益、日峰谷价差套利收益最大化,由深度强化学习模型驱动智能体,寻优调度目标函数,获取最佳电池能量调度时的充放电功率。结果表明,用所提方法调度后,分布式储能电池均在低谷状态中充电、高峰状态中放电,且电池充放电功率在限值之内,配网日网损减少收益、日峰谷价差套利收益增多。
In order to improve the daily network loss reduction benefits and daily peak-to-valley price difference arbitrage incomes of the distribution network,it is necessary to reasonably allocate the charging and discharging power of the battery.Therefore,a distributed energy storage battery energy scheduling method based on deep reinforcement learning was proposed.The remaining available energy estimation model of distributed energy storage battery was constructed,and the battery energy scheduling model based on deep reinforcement learning was used to set the objective function as the daily network loss reduction income and the daily peak-to-valley price difference arbitrage income maximization.The deep reinforcement learning model drove the agent to find the optimal scheduling objective function to obtain the charging and discharging power during the optimal battery energy scheduling.The results showed that after the scheduling by the proposed method,the distributed energy storage batteries were charged in the low state and discharge in the peak state,and the battery charging and discharging power was within the limit,and the daily network loss reduction income of the distribution network and the arbitrage income of the daily peak-to-valley price difference increased.
作者
何山
赵宇明
HE Shan;ZHAO Yuming(Shenzhen Power Supply Bureau Co.,Ltd.,Shenzhen 518001,Cuangdong China)
出处
《粘接》
CAS
2024年第2期193-196,共4页
Adhesion
基金
南方电网科技项目资金资助项目(项目编号:0002200000085094)。
关键词
深度
强化学习
分布式
储能
电池
能量调度
depth
strengthen learning
distributed
energy storage
battery
energy dispatch
