Aggregate solar variability

Under future high solar power penetration, variability in the solar `fuel´ for solar power plants is often thought to cause large short-term voltage fluctuations on distribution feeders and intra-hour load following requirements in balancing areas. Solar power output from 86 photovoltaic (PV) systems in San Diego Gas & Electric territory, the 48 MW Sempra Generation Copper Mountain, and six pyranometers on the UC San Diego microgrid are analyzed to quantify actual variability. The correlation in power output between sites at various distances, aggregate power output variability, and worst-case variability occurrences are analyzed. The variability reduction factor was shown to depend on the ratio of distance between sites and timescale of interest. Consequently, given the size of typical distribution feeders, power output variability on time scales less than 10 minutes is independent, and aggregate variability only grows with the square root of the number of sites. A wavelet variability model (WVM) for simulating solar PV powerplant output given a single irradiance point sensor timeseries was calibrated using the data. The WVM allows studying scenarios for aggregate variability using different feeder and PV configurations. Variability at longer (hourly) time scales is more correlated and can lead to large changes in aggregate power output over one hour. The effects of specific meteorological conditions, such as coastal marine layer clouds on occurrence of large ramps are investigated for coastal Southern California.