Method of classification to assess the variability of horizontal and tilted solar irradiance ramps in a desert climate

Solar irradiance ramps are large changes in solar irradiance over short intervals, from seconds to a few minutes. For non-inertial power generation systems such as photovoltaics (PV), these changes translate to equally rapid output power fluctuations that can have important effects on the system; balancing, in particular, is a critical aspect of the management of such systems, making the prediction and classification of irradiance ramp events essential. In this contribution, we study one-minute solar irradiances in pairs of consecutive minutes, defining ramps as the irradiance change from one minute to the next, through one full year in Doha, Qatar. This arid desert location is characterised generally by low cloudiness throughout the year, yet it experiences high variability in aerosol loads. Three different irradiance measurements are analysed: global horizontal, and global at two south-looking tilt angles, namely 25 and 51 degrees. Ramp occurrences are studied, as well as classifications by ramp type (upward, downward, both) and intensity based on two criteria: a standard deviation score and a “ramp intensity factor”. Ramps are observed only around 2–3% of the time, but are sometimes quite large, reaching around 750 W/m²; on the months with higher irradiance, ramp intensities are lower. The z-score and the Ramp Intensity Factor (RIF) classifications have similar results but RIF provides better view of subtle differences. Finally, an example application of the methodology is shown for evaluation of a nowcasting dataset, providing an additional tool for evaluating model performance beyond standard statistical metrics.