Robust fractional-order adaptive cascaded strategy for mitigating load frequency deviations in renewable thermal hybrid systems

In modern power systems, the increasing share of renewable energy sources (RES) is primarily due to their environmental benefits and sustainability. However, because RES like wind and solar are inherently variable and intermittent besides interfacing with the grid mainly through power electronic devices, they act mainly as power injectors rather than conventional synchronous generators. This results in a reduction of system inertia, intensifying frequency stability challenges. To address this problem, this paper proposes a robust cascaded control strategy for mitigating load frequency deviations in a two-area hybrid renewable-thermal power system. The proposed controller integrates a fractional-order adaptive PID (FOAPID) with a harmony search-optimized PIDA (HS-PIDA), combining the adaptability of Caputo fractional derivatives and model reference adaptive control with the accelerated derivative action of PIDA to enhance dynamic performance. The system model incorporates solar PV, wind, and conventional thermal units, capturing nonlinearities and renewable variability. Six case studies are performed to assess robustness against step and dynamic load changes, renewable intermittency, and parameter uncertainties. Results reveal that the proposed controller reduces settling time for frequency deviations to as low as 3.1 s compared to 14-18 s for TLBO PIDA, SCA PIDA, and Fuzzy PID. Undershoot values are decreased by more than 80% across all cases, while steady-state errors are eliminated where competing controllers failed to converge. Error indices confirm superiority, with SSE reductions of up to 99% and RMSE reductions of up to 97% compared to alternative controllers. Against Marine Predator Algorithm (MPA)-Cascaded PIDA, the proposed scheme further improves overshoot and undershoot by up to 89.85% and achieves faster convergence, proving its effectiveness under diverse operating conditions.