ABSTRACT
This chapter describes why predictions of variability are important and how variability can be better predicted. One method commonly used to measure variability is to construct a histogram of the step changes in power output over time. It discusses the importance of estimating variability and methods to improve the forecasting of variability and the uncertainty associated with that forecast. A complementary method, estimating the power spectral density (PSD), characterizes variability using power spectrum analysis. In a hybrid approach, numerical weather prediction (NWP) models are run every 3 hours, with statistical models used to predict variability at shorter time scales. The variability of solar photovoltaic (PV) power output depends on sunshine intensity and also on the season, time of day, and short-term weather. Grid operators have for many years used load forecasts to reduce the costs that arise from uncertainty in the demand for electricity.
