Deep learning for optimal dispatch of automatic generation control in a wind farm

As a wind farm participates in automatic generation control (AGC), it should trace the real-time AGC signal from independent system operator. To achieve high responding performance, be rapidly distributed to multiple turbines (WTs) via an optimal dispatch. It is essentially non-linear complex optimization due wake effect between different WTs. solve this problem, deep learning employed generate dispatch scheme of farm. The training data acquired results anticipated tasks by genetic algorithm. In order guarantee reliable on-line decision learning, error regulation power command corrected adjustment method rotor speed and pitch angle for each WT. effectiveness proposed technique evaluated compared with methods.