Full Description

Infiltration is a significant factor in a home's energy balance, yet prediction of infiltration flow rates is the least accurate of all the calculations that are used. There are several reasons for this inaccuracy, most of which stem from variations in the construction of houses and in the tightness of fit of the many building components. Infiltration characteristics of a particular structure can be found by several experimental means, one of which is to use a fan pressurization device (FPD), or the so-called "blower door". FPDs are used in many applications, but most commonly in energy conservation retrofit programs. Using the flow rate and pressure differential measurements from the FPD, a natural infiltration rate can be estimated from calculation procedures developed by Lawrence Berkeley Laboratories and elsewhere. The extent to which retrofit measures are recommended is based on the amount of reduction in air leakage indicated by pre- and post-retrofit leak tests. There are several different manufacturers of FPDs and they use different designs and materials in their construction, so there is much uncertainty about their accuracy, reproducibility, and the impact of the operator on the results. This test utilized four different FPDs and three different operators to conduct round robin tests on four different houses according to the appropriate ASTM test standard (ASTM, 1987). The results were analyzed using ASTM E691 (ASTM, 1979) for conducting an interlaboratory test program. The results indicate that subtle errors in fan calibration or operator technique are greatly exaggerated when extrapolating the pressure versus flow curve out to 4 Pa. The total range of results were as much as +/- 40% at 4 Pa, with operator differences (between replications) normally +/- 5% or less. In most cases, the variability of the FPDs themselves was far greater than the variability introduced by using different operators. Variances were somewhat larger for pressurization tests than for depressurization tests. The higher the house pressure used in the computations, the more uniform the test results became. Repeatability variances at 50 Pa were usually better than +/- 5% while the operator differences were usually better than +/- 2%. Combining pressurization and depressurization tests did improve the overall variability slightly. Errors were generally larger on a percentage basis for the tighter houses, since the random errors in equipment components and operator technique constituted a larger fraction of the measurable quantity. The tighter houses required a flow restriction plate to be placed on the fans for the low pressure readings, and fan calibrations indicated a significant difference in calibration results between the open flow and restricted flow cases for two of the three FPDs that used them. In general, the FPD performance could be explained on the basis of their fan calibration curves. The leakage of air around the FPD while it is installed was estimated to produce a 2% error in the worst case, and less than 1% error for the tightest unit.