Experimental validation of an E-GOA-Based MPPT for PV-SEPIC systems under challenging irradiance conditions

In photovoltaic systems, partial shading and rapidly changing irradiance create multiple peaks in the power-voltage (P-V) characteristic, making global maximum power point tracking (GMPPT) challenging. The grasshopper optimization algorithm (GOA) provides a population-based search mechanism for GMPPT but remains sensitive to search boundaries and algorithm parameters. This work establishes, through simulation and hardware experiments, an enhanced GOA (E-GOA) for MPPT of PV-SEPIC systems. In the proposed implementation, the search space is defined directly by the converter duty cycle, with each search agent corresponding to a candidate operating point within the prescribed limits. The search process is reinitialized when the relative variation in PV output power exceeds a predefined threshold. The search dynamics are re-tuned through the social-force parameters and a linearly decreasing control coefficient to balance exploration and exploitation. E-GOA was evaluated in MATLAB/Simulink using a 162 W PV array under ten uniform irradiance levels, ten mismatched patterns and four dynamic scenarios. The methods were benchmarked against PSO, GWO, FA and P&O over 1000 independent runs using tracking efficiency, RMSE, MAE and STD. E-GOA tracked the MPP at 1000 W/m 2 in 0.33s, compared with 0.55 s for FA, 0.64 s for PSO and 0.91 s for GWO. Under mismatched irradiance, E-GOA achieved average tracking efficiencies of 94.21-97.25%, compared with 86.65-95.89% for PSO and 53.46-85.21% for P&O. Experimental validation showed that E-GOA achieved tracking efficiencies ranging from 98.99% to 99.95% across the tested scenarios. The measurement uncertainty of the experimental acquisition chain was also quantified. These results confirm the effectiveness of E-GOA for fast and reliable GMPPT.