Elucidating uncertainty in bifacial photovoltaic gain estimation

• Bifacial gain is estimated with four different monofacial references in Norway. • Bifacial gain estimates depend on the properties of the monofacial reference. • Correcting for irradiance and temperature narrows the range between the estimates. • The best estimate of the bifacial gain is determined to 14.4 %. • Remaining variations in estimates may reflect measurement uncertainty in the method. The bifacial gain of fielded bifacial photovoltaic (PV) systems is commonly estimated by comparing the energy yield of the system with that of a monofacial reference. However, this approach is dependent on the selected reference, and can thus be associated with significant uncertainty, since the two technologies might have very different properties. This study investigates the significance of such differences when the bifacial gain is estimated from energy yield time series data. A series of corrections are applied to account for differences in installed capacity and responses with respect to temperature and irradiance. The analysis is carried out using data collected over two years from a 22.6 kW p ground-mounted fixed-tilt PV system in Norway using four different monofacial technologies as reference systems that all rely on Si as the active material but have different cell architectures. With no corrections applied, the estimated bifacial gain with the monofacial technology that most closely resembles the bifacial is determined to 12.6 % for the two-year period considered. The estimate increases to 13.1 % when correcting for irradiance and 14.4 % when also correcting for temperature. The results demonstrate that the applied corrections reduces the range between the smallest and largest estimate of the bifacial gain from 6.3 to 3.5 percentage points. Although it is possible that the range can be reduced further if differences in spectral response and degradation are accounted for, it should also be expected that some of the remaining variation reflects the measurement uncertainty inherent to this method of estimating the bifacial gain.