Comparative assessment of single-diode model parameter extraction and physical validity for high-efficiency photovoltaic modules

The single-diode model (SDM) is widely used for representing the I – V characteristics of photovoltaic modules. Previous studies have shown that different parameter sets, even non-physical ones, can reproduce the measured curve with low error values. However, physically reasonable parameter sets remain preferable, and few studies have combined grid-based parameter exploration, physical-validity screening, and direct comparison of extraction-formulation-dependent irradiance trends for recent high-efficiency commercial modules. This work investigates commercial high-efficiency PERC and TOPCon modules under 22 irradiance–temperature conditions, comparing a Phang-type local-slope analytical method and a Villalva-type maximum-power-point-matching method through grid-based parameter sweeps, normalized root-mean-square error (NRMSE) evaluation, and physical-validity screening. We find that when a one-diode lumped-circuit model is fitted to illuminated I – V curves under weak irradiance, the extracted ideality factor ( n ) can fall below unity. Such values should not be directly compared with intrinsic dark-diode ideality factors, because the SDM ideality factor extracted from illuminated I – V curves is coupled with the photogenerated current and other lumped-circuit parameters. Non-physical parameter sets were obtained under some sweep conditions and still yielded low NRMSE. The comparison also shows that apparent irradiance trends are method-dependent: the Phang-type method yields irradiance-dependent n and J 0 = J 0 ( G , T ) , whereas the Villalva-type method, with prescribed n , gives an approximately temperature-dominated J 0 ≈ J 0 ( T ) at fixed n . These results indicate that SDM parameters for high-efficiency PV modules are coupled effective fitting quantities, and that error-metric-based validation should be accompanied by physical-validity screening.