在风储配置给定前提下,研究风电与储能系统如何有机合作的问题。核心在于风电与储能组成混合系统参与电力交易,通过合作提升其市场竞争的能力。针对现有研究的不足,在具有过程化样本的前提下,引入强化学习算法。所建立的控制器具备在线学习能力,在学习过程中不断以混合系统收益为反馈信息逐步具备对储能系统充/放电功率、购买备用容量的决策能力。伴随学习时间的累积,将渐进趋于最佳策略,减轻电网调控负担的同时,提高风储合作效率。
Under premise of given wind power and energy storage configuration, this paper studies how wind power and energy storage can cooperate effectively. The core is that wind power and energy storage form a hybrid system to participate in power trading and to enhance its market competition ability through cooperation. Aiming at shortcomings of existing research, reinforcement learning algorithm is introduced under premise of given process samples. Established controller has on-line learning ability, and gain of hybrid system is taken as feedback information in learning process. The controller gradually has the ability to make decisions on power charge/discharge of energy storage system and purchasing reserve capacity. Along with accumulation of learning time, control strategy will gradually become the best one for reducing burden of power system regulation and improving cooperation efficiency of wind power and energy storage.