The development of ultra-large-scale energy storage system(ESS) is beneficial to integrate the real-time renewable energy generation with uncertainty and intermittent features and provide effective capacity support for the power grid. Since renewable energy has strongly decentralized feature, it is effective way to develop distributed dispatch strategy for storage devices to integrate them. This paper focuses on built a multi-timescale dispatch model for ESS. The goal of the proposed model is to minimize the day-aheadoperation cost with better charging/discharging operation allocation for each ESS in each dispatch interval during one day, and smooth the real-time fluctuation s of renewable energy. To address this issue, this paper presents a distributed multi-timescale and discrete-time-gradient optimization algorithm by combining consensus and optimization theories. By using this method, each ESS can response to the change of renewable generation and obtain the optimal operation by using local information share, resulting in better flexibility, robustness, scalability and privacy, etc. Thus, it is suitable for distributed model development and optimal dispatch for ultra-large-scale ESS. Finally, the paper is tested on a system with eleven ESSs and five renewable energy generations. The simulation results show the feasibility and effectiveness of the proposed model and algorithm.
展开▼
