Frequency Regulation of Nonlinear Power Systems using Neural Network Observer-Based Optimized Resilient Controller

This study introduces a resilient frequency controller for nonlinear interconnected power systems to counteract endogenous/exogenous system disturbances. A neural network-based observer (NNO) is intended to estimate lumped system disturbances, such as unmodelled dynamics and unknown disturbances. The estimated NNO’s output is incorporated with a second-order sliding mode controller (SOSMC) to minimize chattering in the control effort and improve the nominal performance of the undertaken plant. The design parameters of SOSMC have been optimally identified by applying Harris hawk optimization (HHO), exercising integral error-based objective function. HHO has demonstrated superior tuning capabilities than other well-known optimization methodologies in terms of convergence rate and transient measurements of system outputs. The asymptotic convergence of estimated error and overall stability of the system has been established employing the Lyapunov argument. System outputs are compared with the results reported in the literature to validate the efficacy of the proposed resilient frequency controller. Presented results showcase the mastery of the applied NNO-based SOSMC over its counterparts in weaker chattering, fast disturbance rejection, and a high degree of robustness against endogenous/exogenous disturbances.