基于数据驱动状态空间映射的风电场下垂-惯性频率调节可行域评估

Increasing penetration of renewable resources integrated into power systems has necessitated frequency support from wind farms. Considering the complex coordination of wind turbines (WTs) and time-variant characteristics of frequency response, it is necessary to assess the primacy frequency regulation (PFR) capability and construct feasible region for wind farms. To cope with the challenges of incomplete parameters in physical-based methods and inadequate interpretability in existing data-driven methods, this paper develops a data-driven state space mapping method for wind farms to construct feasible region of droop and inertia coefficients, and the optimal coefficients are determined by using an iterative scheme. Based on the Koopman operator theory (KOT), nonlinear model of PFR is transformed into a linear lift-dimension form, which can implement assessment with measurement data in real-time. The simulation results validate that the proposed method has the advantages of fast solving, independence from physical parameters, and lower requirements of training data.