Coordinated Planning of Wind Power Generation and Energy Storage With Decision-Dependent Uncertainty Induced by Spatial Correlation

作     者：Yin, Wenqian Li, Yujia Hou, Jiazuo Miao, Miao Hou, Yunhe 

作者机构：Univ Hong Kong Dept Elect & Elect Engn Hong Kong Peoples R China HKU Shenzhen Inst Res & Innovat Shenzhen 518053 Peoples R China 

出 版 物：《IEEE SYSTEMS JOURNAL》 (IEEE Syst. J.)

年 卷 期：2023年第17卷第2期

页      面：2247-2258页

核心收录：

学科分类：0810[工学-信息与通信工程] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 0808[工学-电气工程] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Joint Research Fund in Smart Grid Grant [U1966601] National Natural Science Foundation of China State Grid Corporation of China 

主　　题：Wind power generation Uncertainty Wind farms Correlation Planning Renewable energy sources Generators Decision-dependent uncertainty (DDU) energy storage Gaussian mixture model (GMM) spatial correlation stochastic programming wind power generation planning 

摘      要：The global experience on wind farm development reveals that due to the spatial correlation, the prediction error of wind power is related to the scale of wind farms. This evidence indicates that the uncertainty features of wind power output from large-scale wind farms are not fixed but dependent on expansion decisions. The decision-dependent uncertainty (DDU) will alter the traditional optimization process and pose solution challenges. This article proposes a coordinated planning model for large-scale wind farms and energy storage considering DDU. First, a DDU model, which quantifies the relationship between wind power prediction errors and the wind farm size, is established based on historical data. The proposed DDU model for a single wind farm is extended to multiple wind farms with their spatial correlation captured by a Gaussian Mixture Model. Then, tackling the coupling relationship between decisions and uncertainty, an affine function-based solution method for the stochastic model with decision-dependent probability distributions is proposed. The constructed affine function maps planning decisions to decision-dependent wind power scenario sets via linear transformation. The difference between the planning model with and without the DDU in wind power is compared and discussed. Case studies verify the proposed model and solution method.