UC San Diego

Global solar photovoltaic capacity increased by 35% from 2013 to 2014, and this upward growth trend is likely to continue. Power grids must adapt to accommodate increasing shares of renewable energy penetration. The impact of increasing solar penetration is quantified in terms of the variability and the predictability of net load behavior. As expected, due to variable nature of solar technologies, the predictability of net load decreases with increasing penetration. The need for novel net load forecasting techniques that allow for improved management of grids with high solar penetration is discussed. Integrated net load forecasting methods (solar power forecasts are used as inputs) are recommended for grid operators and utilities. Analysis of forecast performance reveals that the solar variability plays a dominant role in driving the forecasting errors, even more so than the penetration levels. Net load and solar forecast errors are found to be co-integrated, sharing a common stochastic drift. Thus, the solar irradiance time series is sufficient to provide necessary information for the future planning of reserve allocation and storage design for power grids. The benefits of proposed techniques are presented for real- time energy imbalance markets. Design variables regulating the electricity markets and grid timelines govern the system dynamics, which in turn highlight the benefits of forecasting. Increased flexibility of operations at shorter time-scales emerges as a key factor for the reliable and efficient management of power grids