Capacity optimization method for renewable energy joint energy storage based on an edge-lightweight dynamic game

Engineering notes

Engineering notes will be added by the Power for Solar editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为光伏效率、钙钛矿太阳能电池、储能技术、太阳能热利用、BIPV、并网技术等高价值论文补充中文说明。

Original abstract

To address the problems of large renewable output fluctuations, redundant independent energy-storage allocation, and the difficulty of coordinating shared energy storage capacity configuration with service pricing under high wind and photovoltaic penetration, this paper proposes a renewable energy joint energy storage capacity optimization method based on an edge-lightweight dynamic game. First, a joint shared energy storage system architecture consisting of renewable energy station clusters, a shared energy storage operator, a grid dispatch center, and edge control nodes is constructed, and capacity configuration, service pricing, and operation scheduling are incorporated into a unified optimization framework. Second, to overcome the limitation that traditional static games cannot characterize cross-period multi-agent interactions, a dynamic Stackelberg master-slave game model is established, in which the shared energy storage operator acts as the leader and renewable energy stations act as followers, so as to realize closed-loop updates of capacity configuration, pricing decisions, and storage service responses. Third, considering the unknown distribution, limited samples, and strong volatility of wind and photovoltaic forecast errors, a moment-based distributionally robust dynamic chance-constrained model is constructed, and the uncertainty constraints are transformed into second-order-cone tractable forms by using Chebyshev's inequality. Finally, an edge-side lightweight nested second-order cone programming (SOCP) solution method is proposed. Through rolling horizons, warm starts, station-level parallel decomposition, and a lightweight capacity-price update mechanism, the real-time solution capability of the model on low-computing-power edge nodes is improved. Case-study results show that the proposed method can effectively reduce redundant energy storage allocation and system operating costs, improve renewable energy accommodation and storage utilization, and maintain a low constraint violation rate under strong wind and solar uncertainty. Compared with the independent storage allocation method, the proposed method reduces the total system cost by 19.29%, decreases the wind and solar curtailment rate to 2.2%, lowers the average grid-tie deviation to 2.1 MW, and controls the average solution time within 8.6 s, verifying its comprehensive advantages in economy, robustness, and edge-side real-time performance.

Links and sources

Comments