Facilitating wide-band oscillation analysis in wind farms with a novel linearization analysis framework based on the average-value model

Abstract

Wide-band oscillations have become a significant issue limiting the development of wind power. Both large-signal and small-signal analyses require extensive model derivation. Moreover, the large number and high order of wind turbines have driven the development of simplified models, whose applicability remains controversial. In this paper, a wide-band oscillation analysis method based on the average-value model (AVM) is proposed for wind farms (WFs). A novel linearization analysis framework is developed, leveraging the continuous-time characteristics of the AVM and MATLAB/Simulink's built-in linearization tools. This significantly reduces modeling complexity and computational costs while maintaining model fidelity. Additionally, an object-based initial value estimation method of state variables is introduced, which, when combined with steady-state point-solving tools, greatly reduces the computational effort required for equilibrium point solving in batch linearization analysis. The proposed method is validated in both doubly fed induction generator (DFIG)-based and permanent magnet synchronous generator (PMSG)-based WFs. Furthermore, a comprehensive analysis is conducted for the first time to examine the impact of the machine-side system on the system stability of the non-fully controlled PMSG-based WF.