Source-Load Collaborative Frequency Control With Real-Time Optimality in Power Networks

With the increasing uncertainties on both supply and demand sides, how to improve the performance of the current multiple time-scales frequency control architecture is the main focus of this study. Inspired by reverse- and forward-engineering, we propose a novel frequency control scheme based on existing distributed control mechanisms to restore the frequency after disturbances while achieving global economic efficiency. We bridge the time gap between conventional secondary frequency control and economic dispatch. On the load side, after showing that the physical network dynamics can solve well-defined optimal load control problems, we develop a fully decentralized frequency controller that realizes system-wide economic objectives solely based on local measurement. On the generation side, we design a distributed control mechanism so that each bus can automatically track the most economically efficient operating points while realizing frequency recovery. A substitution method is proposed to remove disturbances from the update procedure of auxiliary variables. We characterize the equilibrium and establish the asymptotical convergence of the overall system based on optimization techniques. Finally, a numerical test is conducted to demonstrate the effectiveness of the proposed control scheme.