An Enhanced Polar Light Optimizer with SD-RFAT refinement for efficient parameter estimation of solar PV models

Accurate parameter estimation of photovoltaic (PV) models is essential for reliable performance prediction, system optimization, and fault diagnosis under varying operating conditions. However, conventional analytical, deterministic, and metaheuristic approaches often suffer from premature convergence, sensitivity to initialization, or insufficient numerical stability when applied to highly nonlinear PV models. This study proposes an Enhanced Polar Light Optimizer integrated with a Single-Derivative Newton–Raphson scheme incorporating a Relaxation Factor and Adaptive Tolerance (PLO3–SD-RFAT) for robust PV parameter extraction. The method combines chaotic logistic mapping to enhance global search diversity with adaptive deterministic refinement for stabilized local convergence. The proposed framework is validated using experimental I–V datasets for the RTC France Si solar cell and the Photowatt PWP-201 module under Standard Test Conditions, using Single-, Double-, and Three-diode models. Statistical evaluation over 30 independent runs demonstrates that PLO3 achieves the lowest mean RMSE (3.665×10⁻⁴ for the RTC France TDM case), yielding approximately 0.14% improvement over the closest competitor and up to 18%–69% improvement over lower-performing methods. Additionally, the SD-RFAT refinement reduces computational effort by 15%–25% compared to double-derivative Newton approaches and achieves the lowest convergence variance among tested algorithms. The results confirm that the proposed hybrid framework provides statistically significant improvements in accuracy, stability, and computational efficiency, making it a reliable solution for PV model parameter estimation.