Optimization Strategy of PID-Inspired Metaheuristic Algorithm-Tuned Fractional-Order Controllers Applied to Nonlinear AVR System

The growing penetration of renewable-energy sources into contemporary power networks introduces substantial difficulties in maintaining satisfactory voltage regulation and dynamic stability. Existing PID-based designs for automatic voltage regulators (AVRs) equipped with an amplifier feedback with dynamic weighting (AFDW) architecture provide limited tuning flexibility, and under the considered asymmetric saturation constraints exhibit residual steady-state error (≈0.05 p.u.) and relatively slow settling (>2.2 s), while the previously reported FOPID variant exhibits overshoot above 4%. Moreover, tilt-based structures and comprehensive robustness assessments remain relatively unexplored for this configuration. To address these limitations, a nonlinear AVR with an AFDW architecture and IEEE DC1C saturation limits is developed. The parameters of fractional-order PID (FOPID) and tilt-integral-derivative (FOTID) controllers are optimized under nonlinear saturation constraints using two mathematically inspired metaheuristic approaches: the PID-based search algorithm (PSA) and its recently proposed fractional-order extension, FoPSA. Controller tuning is carried out through a cost function that captures the transient characteristics of the closed-loop response. Extensive simulation studies indicate that FoPSA-based tuning yields more favorable transient behavior for the fractional-order controllers, with notable reductions in settling time and steady-state error relative to the reference designs. In addition, multi-domain evaluations reveal that the FoPSA-FOTID configuration maintains strong robustness against parameter variations, severe load disturbances, and changes in reference-tracking conditions. Overall, the results demonstrate the effectiveness of the proposed control framework and its potential to provide reliable and resilient voltage regulation for modern power systems.