摘要
随着可再生能源产业不断壮大,可再生能源消纳问题日益凸显。为了充分挖掘源–网–荷–储联合运行的灵活性调节能力,提高可再生能源的消纳水平,该文提出一种基于灵活性平衡理论的可再生能源消纳策略。通过Nash博弈实现参与消纳的各主体相互利益达到均衡,并提出一种多智能体迁移强化学习算法。该算法采用了多种人工智能技术,包括基于Nash-Q学习的强化学习技术、资格迹更新技术和迁移学习技术,使学习方式更加灵活、效用更广泛、泛化能力更强。通过算例仿真验证了所提算法在保证最优解质量的同时,具有快速求解的能力,非常适用于求解多主体参与可再生能源消纳问题。
With the growing renewable energy industry, the renewable energy consumption has become a prominent challenge attracting worldwide concerns. A deployment strategy based on equilibrium theory was proposed to fully exploit the flexible operation between source, network, load and storage to improve the consumption level of renewable energy. A multi-agent transfer reinforcement learning algorithm was proposed, in which multi-agents achieve the balance of mutual benefit through Nash game. The algorithm adopted a variety of artificial intelligence technologies, including reinforcement learning technology based on Nash-Q learning, eligibility trace update technology and transfer learning technology, to realize a flexible, effective and generalized learning mechanism with the aim of improving the consumption level. The simulation results show that the proposed algorithm is effective for obtaining optimal solution with fast convergence speed, which is suitable for solving the problem of multi-agent participation in improving the renewable energy consumption level.
作者
李宏仲
王磊
林冬
张雪莹
LI Hongzhong;WANG Lei;LIN Dong;ZHANG Xueying(School of Electrical Engineering, Shanghai University of Electric Power, Yangpu District, Shanghai 200090, China;NARI Research Institute, NARI Technology Co. Ltd., Nanjing 210003, Jiangsu Province, China;Guangdong Power Grid Co. Ltd.,Guangzhou 510600, Guangdong Province, China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2019年第14期4135-4149,共15页
Proceedings of the CSEE
基金
国家自然科学基金项目(51777126)~~
关键词
可再生能源消纳
灵活性
NASH均衡
强化学习
迁移学习
renewable energy consumption
flexibility
Nash equilibrium
reinforcement learning
transfer learning
