reinforcement learning-tuned fractional-order sliding mode control for load frequency stability in power systems

Power systems can improve their frequency stability by using the load-frequency control (lfc) system. power system parameter unpredictability and unforeseen load disruptions complicate and test the lfc system's performance. the goal of this work is to increase the frequency stability of a two-area power system by constructing the fractional-order sliding mode controller (fosmc) inside each area's lfc structure. the controller is combined with reinforcement learning (rl) to further enhance the fosmc's performance against disturbances and parameter uncertainty. the primary objective of this control strategy is to enhance dynamic performance and reduce frequency oscillations in the face of unforeseen load interruptions and uncertainty in all power system components. the performance of the smc is improved by utilizing fractional derivatives in the sliding surface, which successfully reduces chattering and raises the frequency stability of the two-area power system. the performance of the suggested method is assessed in the context of lfc for power systems under various situations by contrasting it with alternative control strategies, such as pdsmc, smc and fuzzy-rl. the findings show that the suggested method (fosmc-rl) significantly improved the frequency stability of the two-area power system. additionally, the suggested approach shows resilience to high power system parameter uncertainty and severe abrupt load disruptions.