Data-driven multi period operation optimization of active distribution networks
Tong Xiaoning, Yueqiang Wang, Qiu Zhangquan, Huang Yang · 2023
Traditional physical models are difficult to model and analyze the uncertainty factors of the power grid, and have weak adaptability to changes in ADN structure and new elements. This paper proposes a data-driven active distribution network operation optimization strategy, which uses data-driven replacement process simulation to directly obtain control strategies from real-time environmental operation information. The implicit internal logic between the optimal operation of the operation scenario and the state of the control equipment can be explored through extreme learning machine, so the mapping relationship between the operation control strategy and distribution network operation scenario can be established. The data-driven reactive power optimization method proposed in this article does not rely on actual system power flow calculations, achieving optimized scheduling of active distribution networks under data-driven conditions, and providing a foundation for control strategies of active distribution networks.