Distributed model predictive control for active power control of wind farm

Abstract

This paper presents the active power control of a wind farm using the Distributed Model Predictive Controller (D-MPC) via dual decomposition. Different from the conventional centralized wind farm control, multiple objectives such as power reference tracking performance and wind turbine load can be considered to achieve a trade-off between them. Additionally, D-MPC is based on communication among the subsystems. Through the interaction among the neighboring subsystems, the global optimization could be achieved, which significantly reduces the computation burden. It is suitable for the modern large-scale wind farm control.