Author(s): Mai Shi, University of Michigan Ann Arbor (First Author, Presenting Author) Michael Craig, University of Michigan Ziqi Wei, Massachusetts Institute of Technology
We examined how two major forces, climate change and the wider adoption of heat pumps and improved insulation, together drive the peak hourly electricity use on US power grids through 2060. By combining hundreds of climate model simulations with a range of heat pump uptake and insulation upgrade scenarios, and applying machine learning to predict hourly demand, we show that considering both factors at once yields peak demands that differ greatly while are far more uncertain than when each is treated alone. In some regions, uncertainty about heat pump adoption explains up to 80 percent of total peak‑load variance, while in others it is climate variability that dominates. This integrated approach can help utilities and planners choose the right investments.
