摘要
目前,风电场配置储能是提高风电经济性较为有效的手段。针对最大化风储合作收益问题,提出一种基于深度强化学习的风储合作决策方法。首先,综合考虑风电、储能系统、外部电网以及需求侧柔性负荷,构建一种新型风储合作系统;然后,针对传统深度强化学习方法易陷入过估计问题,提出改进双竞争深度Q网络(D3QN),并进一步设计基于D3QN的风储合作决策算法;最后,结合实际数据对算法进行仿真验证,结果表明相比传统深度强化学习策略,所提方法能更好协调风电和储能运行,提高风储合作系统的运行收益。
At present,wind farm with energy storage is a relatively effective means to improve the economy of wind power.In order to maximize the benefits of wind-storage cooperative system,wind-storage cooperative decision-making strategy is proposed based on deep reinforcement learning.Firstly,considering wind farms,energy storage systems,external power grids and demand response loads,a new wind-storage cooperative system is proposed.Then,in view of the problem that traditional deep reinforcement learning methods are prone to overestimation,an improved Dueling Double-Deep Q Network(D3QN)is proposed,and a cooperative decision-making algorithm for wind-storage based on D3QN is further designed.Finally,the algorithm is verified by simulation.The results show that compared with the traditional deep reinforcement learning strategy,the proposed method can better coordinate the operation of wind power and energy storage,improve the operating benefits of wind-storage cooperative system.
作者
翟苏巍
李文云
邱振宇
张新怡
侯世玺
ZHAI Suwei;LI Wenyun;QIU Zhenyu;ZHANG Xinyi;HOU Shixi(Electric Power Research Institute of China Southern Power Grid Yunnan Power Grid Co.,Ltd,Kunming 650217,China;Yunnan Power Dispatching Control Center of China Southern Power Grid,Kunming 650011,China;College of IOT Engineering,Hohai University,Nanjing 210098,China)
出处
《智慧电力》
北大核心
2023年第9期60-65,共6页
Smart Power
基金
国家自然科学基金资助项目(62103132)
中国南方电网云南电网有限责任公司科技项目(YNKJXM20220048)。
关键词
风电
储能系统
强化学习
深度神经网络
wind energy
energy storage system
reinforcement learning
deep neural networks
