Frequency regulation feasible region assessment and optimization of wind farm based on data-driven model predictive control

With the integration of large-scale wind turbines (WTs) into grids via electronic interfaces, power system operators have necessitated frequency support from wind farms. Due to the large number of WTs and their complex dynamic characteristics, it is necessary to assess the primary frequency regulation (PFR) capability and construct feasible region of wind farms. In order to cope with the problems of incomplete parameters, analytical solving complexity and the coupling influence of power system regulation characteristics, this paper develops a datadriven state space mapping linear model predictive control (MPC) to assess the maximum PFR capability of wind farms and reasonably distribute coefficients to WTs. Besides, a coordinated iteration framework between dispatching center and wind farms is proposed to further optimize the wind farm regulation feasible region. The simulation results verify that the proposed method has the advantages of independence from physical parameters, fast analytical solution, and lower requirements of training samples on limited scenarios.