Das, Milton Kumar2025-02-192025-04-012025-02-192021-03-16https://doi.org/10.1002/2050-7038.12837https://dspacenew8-imu.refread.com/handle/123456789/2670The inequality between active power generation and load demand due to volatile power generation and random changes of load poses a challenge to maintain the system frequency. To address this, in the present work, reinforced learning neural network based–proportional-integral-derivative (PID-RLNN) controllers are proposed in the discrete mode load frequency control (LFC) problems for a three-area hydrothermal hybrid distributed generation (HDG) deregulated power system. The proposed power system is incorporated with super conducting magnetic energy storage system (SMES), and each control area is interconnected via AC/DC parallel links. The dynamic responses using PID-RLNN controllers are compared with PID controllers and RLNN controllers for different loading conditions. The gains and parameters of all the controllers are optimized using grasshopper optimization algorithm (GhOA). Investigation reveals that the PID-RLNN controllers give better dynamic performances compared to PID and RLNN controllers for different loading conditions. Sensitivity analyses are performed to investigate the robustness of the proposed controllers, and investigation reveals that their performances are quite robust for variation of inertia constant and loading conditions. The fluctuation of loads in each area is also considered to analyze the performances of proposed controllers and it is seen that the proposed PID-RLNN controllers damp the oscillations of dynamic responses.enArticle