一种基于局部到全局映射的具有通信容错能力的风电场分散控制方法

This article presents a modular decentralized control method for wind farms (WFs) to optimize power dispatch among wind turbines (WTs) without instant communication. By harnessing the nonlinear fitting capability of two data-driven modules, this method facilitates decentralized control of each WT with existing centralized control methods. Specifically, the local-to-global mapping module captures the intricate relationship between the historical local state variables of each WT and the current global state variables of the WF. The open-loop and closed-loop predictive modes enable the demanded power prediction module to locally predict the active power required by the transmission system operator (TSO) under various communication delays. Two classical multiobjective control modes are incorporated into the local controller depending on whether it requires data from the TSO. Testing on a WF with 32 WTs in MATLAB/Simulink validates that the proposed method has robustness and effectiveness closely aligned with centralized control methods.