Adaptive Control and DC-Link Regulation of Grid-Connected PV-PEMFC Systems Using Dual-Layer MPPT Architecture

In this research, a two-tier control architecture is proposed for a 100 kW grid-connected photovoltaic-proton-exchange-membrane fuel-cell (PV-PEMFC) system that interfaces a 25 kV utility through a DC-DC boost stage and a three-level voltage-source converter. Level 1 performs maximum-power-point tracking (MPPT) with a perturb-and-observe (P&O) algorithm, whereas Level 2 regulates the d-q converter currents. Multiple current-control techniques are investigated in Level 2, including a fuzzy-logic controller (FLC), a genetic-algorithmtuned fractional-order PID (GA-FOPID), and a tunicate-swarm-algorithm-tuned FOPID (TSA-FOPID). Simulation tests subject the hybrid source to rapid 800-1000; W m-2 irradiance change, a 20°C to 30°C temperature swing, and a ramped hydrogen-flow profile. The proposed P&O+FLC strategy delivers the highest MPPT efficiency (-0.7% steady-state error), the lowest power ripple (< 1%), and the tightest DC-link regulation (± 0.09%) while meeting IEEE-1547 harmonic limits. GA-FOPID shortens settling time but incurs overshoots up to 50 V, whereas TSA-FOPID yields low-noise responses at the expense of steady-state power throughput. These results show that, despite the sophistication of evolutionary tuning, a judiciously designed fuzzy P&O scheme offers the best compromise between dynamic response and steadystate quality. The dual-layer architecture therefore constitutes a practical, grid-compliant solution for real-time microgrids and future converter-less PV-PEMFC plants.