Data-driven Modeling of Thermostatically Controlled Loads

The intermittent nature of renewable energy sources such as wind and solar, as well as their ever-increasing penetration, means that the ability of the power grid to respond to generation swings is needed more than ever. Thermostatically controlled loads (TCLs) can significantly contribute to this goal. However, due to the social and temperature dependence of TCLs, obtaining an accurate model for their operation becomes a challenge. In this paper, we use modern data science techniques and data collected from an actual home to extract data-driven TCL models. Such models can be used for home demand forecast or optimal home energy management applications to assist in grid operation.