A slope-adaptive hybrid spline framework for global maximum power point tracking in partially shaded photovoltaic systems

Partial shading can produce multimodal power–voltage ( P – V ) characteristics and cause conventional maximum power point tracking (MPPT) methods to converge to a local maximum. This paper proposes a slope-adaptive hybrid spline (S-AHS) framework that switches between linear and cubic interpolation according to local slope variation and combines the reconstructed P – V curve with a coarse-to-local global search. The revised validation clearly separates measured and simulated information: the uniform-irradiance P – V comparison uses the experimental curve available from the El Jadida photovoltaic station, whereas the dynamic time-series test is a simulated, experimentally informed benchmark rather than a direct hardware experiment. In the original time-series benchmark, S-AHS gives NMSE = 0.0007, R = 0.9935, and FB = 0.0012. To directly test the global-MPPT claim, four additional datasheet-calibrated partial-shading patterns with bypass-diode activation were introduced. They produce three local maxima per P – V curve. With 51 sampled points and α = 1, S-AHS identifies the global peak with tracking efficiencies between 99.78% and 99.98% and a worst-case peak-power error of 0.22%. A sensitivity study over α ∈ [0.5, 1.5] shows a mean peak-power error of about 0.11%, while the mean fraction of intervals treated cubically decreases from 24.5% to 5.5%. In 50 noisy repetitions with a power-noise standard deviation equal to 0.5% of the true GMPP, S-AHS retains mean tracking efficiencies from 99.74% to 99.90% across the tested PSC cases. These results support S-AHS as a computationally economical interpolation-based GMPPT strategy while also clarifying the present study’s limitation: the new broad PSC tests are controlled numerical benchmarks and do not replace multi-technology hardware validation.