Interdependence modeling of wind farm frequency support feasible region: A non-iterative system-wide dynamic characteristics scheduling

The provision of frequency regulation support (FRS) services by wind farms (WFs) is of crucial importance for frequency stability of power systems with high-penetration renewable energy. With time-varying wind speed, real-time scheduling for the FRS characteristics of WFs is essential for ensuring the security of system frequency and dynamic power flow (PF). However, the extensive number of wind turbines (WTs) and interdependence of frequency support capabilities (FSCs) among WFs contribute to the complexity of FRS dynamics, rendering the quantification of FRS security of WTs challenging, especially in the absence of precise WT parameters. Therefore, this paper proposes a data-driven method for modeling interdependence of FSCs across WFs. Utilizing space transformation, the original complex nonlinear FRS dynamics of WTs are transformed into a dimension-augmented linear model, facilitating the construction of an analytical expression for FSCs. On this basis, an optimal scheduling model considering the interdependent characteristics of FSCs is developed, which can be solved by employing a hybrid algorithm combining Kriging-assisted surrogate with piecewise elite learning strategy. The simulation results demonstrate that the proposed method enables fast online scheduling of FRS characteristics for WFs, minimizing FRS costs while maintaining system frequency, WTs, and PF security, and enhances computational efficiency by 98.54 % without reliance on physical parameters. • System-wide scheduling with interdependent dynamics eliminates multiple iterations. • Optimized coefficients ensure security of frequency, turbines, and power flow. • Data-driven dimension-augmented method constructs interdependent feasible regions. • Integrating kriging-assisted agent and elite learning enables efficient solutions.