Unintended consequences of smart thermostats in the transition to electrified heating

Abstract

As building space heating undergoes an increasingly rapid transition toward electrification, it is vital to understand the impacts of these new electrical loads on the grid for future energy resource planning. While current methods for estimating heating demand rely on building modeling and occupant behavioral assumptions, we provide a scalable, data-driven approach for estimating regional electrical demand using real-world data from thousands of homes in a new, publicly available smart thermostat dataset. We find that despite lowering overall energy consumption, smart thermostat control algorithms can severely increase the winter peak heating demand through load synchronization during the early morning hours, when solar energy is unavailable. These peaks present unintended system-level consequences of focusing purely on local energy efficient control and can hinder the integration of renewable energy and electric heating. As a resource for future energy system planning, we provide our methodology as an open-source toolkit that can be used to analyze other regions around the world.