Menu

Building energy codes are used worldwide to promote energy efficiency in buildings. Although these codes are widespread and have the potential to yield high energy savings, few analyses have measured their impact by using real energy consumption data. In this paper, Climate Policy Initiative San Francisco assesses the impact of state energy codes using residential energy use data at the state level. By conducting a regression analysis comparing states with building energy codes to those without, CPI SF measures the realized energy savings of energy codes and compares them to existing estimates based on building simulation models.

Key Findings

The building energy codes studied are associated with:

1. Lower energy consumption per housing unit. We find a decrease of roughly 10% in energy use relative to households that were not built under these codes.

2. A shift toward natural gas and away from lesser-used “other” fuels, most notably fuel oil. We show that housing units built under the studied codes derive a greater share of their energy from natural gas and less from other fuels than units not built under these codes. This finding may reflect provisions in the studied codes that encourage high-efficiency gas units and electric heat pumps.

3. Lower emissions per housing unit. Lower energy use in code buildings reduces energy-related emissions. Moreover, generating energy from direct combustion of natural gas is less greenhouse gas intensive than burning other fuels such as fuel oil. Our results suggest that the combined effect of energy savings and fuel-switching has delivered about a 16% reduction in greenhouse gas emissions from an average code household.

This is the first national study to measure the impact of codes using real household energy use data. Previous studies have focused on specific energy sources such as electricity, or used engineering simulations rather than actual energy consumption data. Our results suggest codes are delivering greater energy savings than the 5% savings engineering models predict.

up

Esta página contém posts em múltiplos idiomas

Usamos cookies para personalizar o conteúdo por idioma preferido e para analisar o tráfego do site. Consulte nossa política de privacidade para obter mais informações.