面向新型电力系统的Transformer与Q学习双驱动智能发电控制OA

[Objectives]The increasing integration of distributed renewable energy with higher uncertainty and fluctuation into power systems leads to greater frequency deviations and exacerbates active power imbalance in the systems.To effectively address the active power imbalance caused by renewable energy integration,this study proposes a Transformer and Q learning dual-driven smart generation control(TQDD)algorithm.[Methods]The smart power generation controller based on the TQDD algorithm consists of a digital-analog dual-drive loop and a proportional loop.Within the digital-analog dual-drive loop,complementary ensemble empirical mode decomposition with adaptive noise(CEEMDAN)performs mode decomposition on the acquired frequency deviation signal.Transformer is used to predict a series of modal components after mode decomposition.K-means clustering classifies the predicted signals into large and small fluctuation signals.Q learning tracks the large fluctuation signals after classification.The fractional-order PID(FOPID)rapidly tracks the small fluctuation signals after classification.[Results]The proposed TQDD method and five other comparative algorithms are simulated in a two-area power system case in which the output proportion of renewable energy generating units is 80%.The results show that under the control of the TQDD algorithm,frequency deviation,total power generation cost,and carbon emission cost are reduced by 45.44%,12.04%,and 10.25%,respectively,compared with other comparative algorithms.[Conclusions]The proposed algorithm enables precise control of the output power from each generating unit in new-type power systems,thereby reducing frequency fluctuations in the power systems.