Tracer gas techniques can be used to measure air infiltration rates in buildings, but it is expensive and time-consuming to obtain a complete characterization of airtightness through these methods. Instead, there are models that predict air infiltration rates from pressurization test results and weather conditions. Other models predict air infiltration without the input of a pressurization test.

In order to study the predictive accuracies of these models, a group of identical homes in Freehold, NJ was subjected to both pressurization and tracer gas measurements of infiltration. The infiltration and pressurization test results are compared. Four air infiltration models are used to predict the infiltration rates in the houses, and these predictions are compared with the measured rates. The predictions are made using several, different values of the inputs required for the models, and the effects of varying these inputs are studied. Use of identical houses eliminates variability in the pressurization-infiltration relation due to house style and vintage. Simultaneous measurements of infiltration in all the houses enable analysis of each night’s data without the confounding effects of weather.