Comparative Assessment of Solar Irradiance Transposition Models for PV Systems in Durban, South Africa: Evidence from 20 Years of Climate Data (2001–2020)

Accurate estimation of solar irradiance on tilted photovoltaic (PV) surfaces is central to performance prediction, system sizing, and financial assessment of solar energy projects. This study conducts a comprehensive comparative evaluation of five widely used irradiance transposition models—Liu & Jordan, Hay & Davies, Klucher, Perez, and Reindl—using 20 years (2001–2020) of climate data for Durban, South Africa. The models were assessed based on their ability to estimate global tilted irradiance (GTI) and the downstream impacts on PV system performance, including energy yield, performance ratio (PR), capacity factor (CF), solar fraction (SF), and avoided CO₂ emissions. A 7.2 kWp rooftop PV system with typical system losses, climatic inputs, and realistic operating assumptions was modelled to quantify differences across models. Results show a 7% variation in annual energy output and CO₂ mitigation potential, driven primarily by differences in diffuse irradiance treatment. Perez and Reindl consistently produced the highest GTI and PV output due to their robust handling of anisotropic sky conditions, while Klucher and Liu & Jordan delivered conservative estimates. Statistical error analysis (MBE, RMSE, nRMSE) confirmed Perez as the benchmark and Reindl as the closest-performing model. Overall, findings highlight the importance of transposition model selection for accurate PV performance evaluation, system bankability, and climate impact projections in diffuse-dominant subtropical climates.