Nonlinear Model Order Reduction Method of PMSG-Based Wind Farm for Large Disturbance

This paper presents a multi trajectory piecewise linear (MTPWL) based method for reducing the order of wind farm for large disturbance. Firstly, the nonlinear system of wind farm is linearized at multiple linearization points along multiple training trajectories, and then the subsystems is reduced by Iterative rational Krylov algorithm (IRKA). Then, the dynamic characteristics of the wind farm nonlinear model are approximated by the weighted combination of low-order linear subsystems. Finally, the modified system with the PMSG-based wind farm is used as the test system. By applying different faults to the test system and comparing the responses obtained by full order model (FOM) with reduced order model (ROM), the effectiveness of the proposed method in capturing the dynamic behavior of PMSG-based wind farm in fault scenarios is verified.