Prediction Differences with Sub-Hourly Weather in Cloudy Climates
Bright sunny days are frequently broken up by intermittent clouds which move across a PV system at a sub-hourly timescale, causing intermittent clipping. Therefore, using hourly data to predict power on bright but cloudy days can lead to an over-prediction if the average irradiance during an hour results in generated power greater than the max output of the inverter. In order to more accurately quantify the difference in predictions with sub-hourly weather, we analyzed several PV systems using SAM, PVsyst, and SolarFarmer and compared the predictions with measured data. The mean bias error was correlated with the mean daily clearness index to derive a correction function for hourly measurements. A regional map of correction factors are provided for the United States, which can be used with hourly data to reduce over-prediction error. For the lowest uncertainty, the use of sub-hourly measurements are recommended when available.