Probabilistic approach for site-adaptation and economic performance estimation of a photovoltaic project in South Korea

Site-adaptation corrects long-term satellite-derived solar irradiance datasets, which improve solar power system modeling by reducing the risks in performance estimation. This study begins with a data-efficient approach to probabilistic site-adaptation, using simple model averaging to quantify uncertainties in the corrected dataset. The proposed model reduces the continuous ranked probability score of up to 34 W/m2 and decreases the bias to nearly 50 % of the initial value. The model performance is associated with the quality of cloud modeling or the seasonal variations of meteorological parameters. In the next phase, the output from the site-adaptation model is employed to develop a probabilistic method for estimating economic performance with a skewed normal distribution. This approach is demonstrated in a case study of a photovoltaic system in South Korea under power purchase agreement (PPA). The evaluation of the distributional economic performance shows that the estimated values are generally aligned with the observational data with a high degree of probability and exhibit greater accuracy than initial estimates. This correction of economic performance estimation is crucial, as evidenced by the demonstration that it can increase the debt service coverage ratio (DSCR) to satisfy a specific requirement (e.g., DSCR >1.3), and turn a negative net present value positive, thereby suggesting a shift in the evaluation of projects from economically infeasible to feasible. Sensitivity analysis shows that the economic performance indexes are highly influenced by the PPA price, debt ratio, and total installation cost.