Independent assessment of temperature models for floating photovoltaics

Floating photovoltaic (FPV) systems may operate at lower temperatures than ground-mounted photovoltaic (GPV) systems due to the cooling influence of the underlying water surface. Numerous studies have proposed various methods for accounting for this when simulating the energy yield of FPV systems. However, studies have reported a wide range of heat loss coefficients, and FPV-specific models yield significantly different results. To address these limitations, this study provides an independent comparison of temperature models using three years of operational data from an FPV installation that has not been used in previous validations. Five FPV temperature models and two GPV models were evaluated. Results show that conventional GPV models perform comparably well for air-cooled, inland FPV systems. The study also concluded that the use of complex analytical models is not justified, as they exhibited lower accuracy and slower computational performance than empirical models. Nevertheless, empirical models should be used with caution as they yield reliable predictions only when applied to systems similar to those on which they were developed. Finally, the study demonstrated that model selection can impact annual DC energy yield estimates by up to 2.3%, underscoring the importance of selecting suitable temperature models and coefficients. • FPV temperature models were benchmarked using three years of operational data. • GPV models perform well for inland, air-cooled floating PV installations. • Large range of PVsyst heat loss coefficients in the literature for similar FPV systems. • Analytical temperature models showed lower accuracy and higher computation times. • Annual FPV energy yield differed by 2.3% depending on the temperature model.