PNNL

Abstract

Projections of future building electricity consumption are used in a variety of fields and for a variety of purposes, from energy utility investment decisions to global climate assessments. Existing approaches to modeling building electricity consumption similarly span a range of structural methodologies, spatial resolutions, and temporal scales, potentially leading to divergent projections. This paper compares how two such models with different structures and resolutions respond to a common set of population and climate drivers in the western U.S. The BEND model simulates hourly residential and commercial building electricity consumption at the county scale by weighting the results of hourly simulations of representative building types. For the region assessed in this study, BEND performs approximately 100,000 unique building simulations. In contrast, the projected electricity demand in GCAM-USA is determined for each state, year, sector (residential, commercial) and service (e.g., heating, cooling, and others), based on population, income, technology, energy prices, and average annual climate, among other factors. This paper compares projections by aggregating the two models’ results to a common resolution: annual residential and commercial building electricity consumption by state. The results show a high degree of consistency in most states. Both models show similar responses to future projected population change and climate change, with population change having the larger impact between the two. State-level annual building electricity consumption appears relatively insensitive to model structure and resolution. Differences are primarily due to how the models capture changes in the aggregate energy efficiency of the building stock as it evolves over time.