Design of a Novel Smart Generation Controller Based on Deep Q Learning for Large-Scale Interconnected Power System

Abstract

This study proposes a novel control strategy based on deep Q learning (DQL) for smart generation control of large-scale interconnected power systems. In this novel DQL algorithm, a deep neural network with higher intelligence is adopted by the action selector of Q learning. Based on the forecasting and online adjustment, DQL could speed up the convergence rate of Q learning, and mitigate the curse of dimensionality. The feasibility and effectiveness of the designed DQL controller are demonstrated in the simulations considering an IEEE two-area load frequency control model and a four-area model based on China Southern Power Grid with 23,328 different parameters. Simulation results verify that the DQL-based controllers can optimize the active power in multiareas and minimize the overall system frequency deviations.