Characterizing the ramps and noise in solar power imbalances

The output of solar photovoltaic resources can be variable, with the power variability manifesting across multiple time scales. Several studies characterize the impacts of variability on system operations but there is limited literature available on how this variability needs to be modeled. This paper decomposes high resolution solar power measurements into metered average, ramp and noise components. With explicit consideration of average values and the imbalance power around the average, the proposed approach is able to consider the dispatch practices which use average values across dispatch intervals while the supply/demand changes continuously. The ramp component captures this disconnect while the noise reflects the natural variations in the solar output due to cloud transitions and other factors. The models proposed here allow us to identify the ramp components even when high frequency measurements are not available. A case study of actual field measurements of solar generation is presented to demonstrate how noise reduces with spatial aggregation while the ramp component persists in terms of magnitude across higher aggregations. A use case for the ramp component for balancing reserve sizing is also presented compared with the conventional one.