Correlating Photovoltaic Clear-Sky and Actual Output to Sky Cover Measurements
David R Daily, David G Holmberg, John S Baras
To appear in the proceedings of 39th IEEE Photovoltaic Specialists Conference, Tampa, Florida, June 16-21, 2013
As building systems grow more complex and electric rates move toward dynamic prices for demand and energy, facility owners stand to gain in energy savings from understanding the expected performance of their building systems. Hourly, forecast energy generation from photovoltaic sources allow owners to alter their generation and consumption for their optimal benefit. This paper presents an empirical approach for mapping weather variables to photovoltaic (PV) system energy output. The resulting correlations allow an estimated forecast PV system output based on a forecast weather variable. In this paper, publicly available sky cover data is correlated to PV output in a two-step process. The first step estimates PV system clear-sky output as a function of date and time based on Sky Cover sensor indications of “clear sky”. The second step correlates actual output to sky cover normalized by the estimated clear-sky output. The accuracy of the correlation is discussed with emphasis on the impact of the sky cover sensor location at a significant distance from the PV array.