Residential energy and IAQ models

Americans spend most of their time inside residences where they are exposed to a number of pollutants of both indoor and outdoor origin. Residential buildings also account for ~20% of the primary energy consumed in the U.S. Energy use and indoor air quality (IAQ) are inextricably linked, as many source and removal mechanisms of indoor pollutants are closely tied to energy consumption, and changes made to improve indoor air quality can impact energy consumption. To provide a tool for future investigations of interactions between energy use and IAQ in homes across the U.S. population, we developed a custom set of nationally representative building energy and IAQ mass balance models that predict annual energy use for space conditioning and indoor concentrations of a number of pollutants of both indoor and outdoor origin across the U.S. residential building stock. The residential energy and indoor air quality (REIAQ) model framework is built in Python and integrates between EnergyPlus and a dynamic mass balance model. REIAQ utilizes historical weather data to predict hourly energy consumption, air change rates, and HVAC system runtimes, which are coupled with historical outdoor pollutant concentration data and assumptions for indoor emission sources and other factors to predict hourly indoor pollutant concentrations. The model set uses 3971 individual home models built from 209 distinct home geometries modeled after a NIST CONTAM database located in 19 U.S. cities.

REIAQ model workflow in Python

The model process involves the following sequential steps: (1) manually constructing a minimal number of typical home geometries required in BEopt, (2) modifying those base home geometries to include region-specific details on envelope construction, HVAC characteristics, and other characteristics for use in energy simulations, (3) running hourly energy simulations in EnergyPlus, (4) passing hourly energy simulation outputs such as modeled hourly air change rates (ACRs) and central HVAC system runtimes to a transient indoor air mass balance model to simulate hourly concentrations of several priority pollutants of both indoor and outdoor origin, and (5) aggregating hourly model results over the course of the model year and applying population-weighting factors to generate nationally representative average concentrations of each pollutant and an aggregate estimate of the total annual heating and cooling energy consumption in U.S. residences. 

Modeled indoor pollutants include PM2.5, UFPs, O3, NO2, and several volatile organic compounds (VOCs) and aldehydes. The REIAQ model set successfully predicted annual space conditioning energy consumption for the U.S. residential building stock within ~2% of historical data. Modeled indoor concentrations, infiltration factors for outdoor contaminants, and indoor/outdoor ratios of each pollutant all matched closely with observations from prior field studies. Population-weighted annual average indoor pollutant concentrations were also used to estimate the chronic health burden of residential indoor exposures.

More details on the REIAQ model set will be included in a forthcoming paper.

Model outputs are available for download below:

This project is funded by the US Environmental Protection Agency (EPA-G2014-STAR-A2).