In 2008, we established a sensor network uniquely capable of studying the solar intermittency of dense urban rooftop PV. Continuous 1-second global horizontal incident irradiance (GHI) data is being collected at 8 stations located throughout the 1200-acre (5 square km) UC San Diego campus in coastal southern California. A Total Sky Imager takes hemispherical sky photographs, and a ceilometer records cloud height. Given an existing one-year dataset of 1-second and 5-minute GHI measurements, statistical properties of extreme ramp rate events for individual and aggregated sites on a virtual distribution feeder were derived. This analysis allows the estimation of benefits of energy storage systems and provides different probabilistic scenarios for input into PV system models. Conditional averaging was employed to study the prelude and following of the largest ramp rate events. Reflection of solar radiation by approaching clouds was found to cause a strong increase in GHI prior to the shaded period, thereby increasing the ramp rate by more than 30%. This analysis provides critical information about how GHI varies during strong ramping events, which will aid the modeling of these phenomena.
Kleissl, J., Harper, J., & Dominguez, A. (2010). A solar resource measurement network for solar intermittency at high spatio-temporal resolution. In Proc. ASES Annual Conf., Phoenix, AZ.