The synthesis of solar radiation data for sizing stand-alone photovoltaic systems

The author describes a model that synthesizes long-term hourly insolation data from monthly average insolation data using statistical models developed from 23.5 years of hourly data from 20 weather stations scattering across the United States. This model was developed specifically for sizing stand-along photovoltaic power systems and was therefore verified by comparing system performances resulting from the synthetic data with performances resulting from actual hourly data. Comparisons of the system performances indicate that the model is valid for system sizing regardless of array orientation, seasonal changes in the load, or the average system loss of energy. In all cases, the relative error between array sizes resulting from actual and synthetic data was less than 6%. Overall, the error was less than 2.5%. The model has one limitation in that it was designed specifically for stand-alone photovoltaic systems with energy storage and should not be applied to systems without storage. In general, it is valid for any system with two or more days of energy storage