SunDance

Abstract

The increasing penetration of grid-tied solar is complicating utilities' ability to balance electricity's real-time supply and demand. While advancements in solar forecast modeling are enabling utilities to better predict and react to future variations in solar power, these models require historical solar generation data for training. Unfortunately, pure solar generation data is often not available, as the vast majority of grid-tied solar deployments are behind the meter, such that utilities only have access to net meter data that represents the sum of each building's solar generation and its energy consumption. To address the problem, we design SunDance, a black box technique for disaggregating solar generation from net meter data that requires only a building's location and a minimal amount of historical net meter data, e.g., as few as two datapoints. SunDance leverages multiple insights into well-known fundamental relationships between location, weather, solar irradiance, and physical deployment characteristics to accurately disaggregate solar generation from net meter data without access to a building's pure solar generation data for training. We also identify and leverage a new fundamental relationship, which we call the Universal Weather-Solar Effect, that, to the best of our knowledge, has not been articulated in the past and is broadly applicable to other solar energy analytics. We evaluate SunDance using net meter data from 100 buildings and show that its black-box approach achieves similar accuracy without access to any solar training data as a fully supervised approach with complete access to such training data.