Blalh  blah

by Alan Meier
Alan Meier is executive editor of Home Energy Magazine.
This Home Energy classic, originally printed in 1986, explains a simple way
to take one air infiltration measurement and determine a home's average air
infiltration rate.
Many researchers have sought to develop a correlation between a onetime
pressurization test and an annual infiltration rate. Translating blower
door measurements into an average infiltration rate has bedeviled the
retrofitter and researcher alike. The rate of air infiltration constantly
varies, yet the pressurization test is typically a single measurement.
Nevertheless, many researchers have sought to develop a correlation between
a onetime pressurization test and an annual infiltration rate.
ACH Divided by 20
In the late 1970s, a simple relation between a onetime pressurization test
and an average infiltration rate grew out of experimentation at Princeton
University. For a few years, the correlation remained "Princeton folklore"
because no real research supported the relationship. In 1982, J. Kronvall
and Andrew Persily compared pressurization tests to infiltration rates
measured with tracergas for groups of houses in New Jersey and Sweden.
They focused on pressurization tests at 50 Pascals because this pressure
was already used by the Swedes and Canadians in their building standards.
(This measurement is typically called "ACH50.") Other countries and groups
within the United States have also adopted ACH as a measure of house
tightness. Persily (now at the National Institute of Science and
Technology) obtained a reasonably good estimate of average infiltration
rates by dividing the air change rates at 50 Pascals by 20, that is:
average infiltration rate (ACH) = ACH50(1)

20
In this formula, ACH50 denotes the hourly air change rate at a pressure
difference of 50 Pascals between inside and outside. Thus, for a house with
15 ACH at 50 Pascals (ACH50 = 15), one would predict an average air change
rate of (15/20 = ) 0.75 ACH.
This simple formula yields surprisingly reasonable average infiltration
estimates, even though it ignores many details
These "details" are described below:
• Stack effect. Rising warm air induces a pressure difference, or "stack
effect," that causes exfiltration through the ceiling and infiltration
at (or below) ground level. The stack effect depends on both the outside
temperature and the height of the building. A colder outside temperature
will cause a stronger stack effect. Thus, given two identically tall
buildings, the one located in a cold climate will have more stackinduced
infiltration. A taller building will also have a larger stack
effect. Even though outside temperature and building height affect
average infiltration rates, neither is measured by the pressure test.
During the summer, stack effects disappear because the inside air is
usually cooler (especially when the air conditioner is operating). Windinduced
pressure therefore becomes the dominant infiltration path.
• Windiness and wind shielding. Wind is usually the major driving force in
infiltration, so it is only reasonable to expect higher infiltration
rates in windy areas. Thus, given two identical buildings, the one
located in a windy location will have more windinduced infiltration.
Nevertheless, a correlation such as ACH50/20 does not include any
adjustment for windiness at the house's location. Trees, shrubs,
neighboring houses, and other materials also shield a house from the
wind's full force. Since a brisk wind can easily develop 10 Pascals on a
windward wall, the extent of shielding can significantly influence total
infiltration. A pressurization test does not directly measure the extent
of shielding (although a house with good shielding may yield more
accurate measurements since it is less affected by wind).
• Type of leaks. The leakage behavior of a hole in the building envelope
varies with the shape of the hole. A long thin crack, for example,
responds less to variations in air pressure than a round hole does. The
pressure/air change curve (determined with a calibrated blower door)
often gives clues to the types of leaks in a house.
A person conducting pressurization tests on a particular house can collect
considerable information about these details. For example, it is easy to
measure a house's height and estimate the wind exposure. The kinds of
cracks can often be judged through careful inspection of the building
construction. Climate data, including windiness and temperature, can be
obtained from local weather stations. Ideally, this additional information
should be applied to the formula in order to get a correlation factor more
accurate for that house. Unfortunately, the formula was developed from data
in just a few houses in New Jersey and Sweden, and it cannot be easily
adjusted to other locations and circumstances. Should a retrofitter in
Texas also use ACH50/20, or is dividing by 15 more appropriate for the
Texas climate and house construction types?
The LBL Infiltration Model
Researchers at Lawrence Berkeley Laboratory developed a model to convert a
series of fan pressurization measurements into an "equivalent leakage
area." (See HE, "Blower Doors: Infiltration Is Where the Action Is,"
Mar/Apr. '86, p.6. and the ASHRAE Book of Fundamentals chapter on
ventilation and infiltration.) The equivalent leakage area roughly
corresponds to the combined area of all the house's leaks.
A second formula converts the equivalent leakage area into an average
infiltration rate in air changes per hour. This formula combines the
physical principles causing infiltration with a few subjective estimates of
building characteristics, to create relatively robust estimates of
infiltration. ASHRAE has approved the technique and describes the formulae
in ASHRAE Fundamentals. The LBL infiltration model is now the most commonly
accepted procedure for estimating infiltration rates.
Max Sherman at LBL used this model to derive the theoretical correlation
between pressure tests at 50 Pascals and annual average infiltration
rates.1 His major contribution was to create a climate factor to reflect
the influence of outside temperature (which determines the stack effect)
and windiness. Sherman estimated the climate factor using climate data for
North America and plotted it (see Figure 1). Since the factor reflects both
temperature and seasonal windiness, a cold, calm location could have the
same climate factor as a warm, windy location. The map also reflects summer
infiltration characteristics. Note how Texas and Vermont have the same
climate factors.
Sherman found that the correlation factor in the revised formula could be
expressed as the product of several factors:
correlation factor, N = C * H * S * L (2)
where:
C = climate factor, a function of annual temperatures and wind (see
Figure 1)
H = height correction factor (see Table 1)
S = wind shielding correction factor (see Table 2)
L = leakiness correction factor (see Table 3)
Values for each of the factors can be selected by consulting Figure 1 and
Tables 13. An estimate of the average annual infiltration rate is thus
given by
average air changes per hour = ACH50 (3)

N
This formula provides a more customized "ruleofthumb" than the original
ACH50/20 , when additional information about the house is available
An Example
The application of the climate correction is best shown in an example.
Suppose you are pressure testing a new, lowenergy house in Rapid City,
South Dakota. It is a twostory house, on an exposed site, with no
surrounding vegetation or nearby houses to protect it from the wind.
1. At 50 Pascals, you determine that the ACH50 is 14.
2. You consult Figure 1, and determine that the house has a climate factor,
"C," of 1417. Since Rapid City is near a higher contour line, select
17.
3. The house is two stories tall, so the appropriate height correction
factor, "H" (from Table 1), is 0.8.
4. The house is very exposed to wind, and there are no neighboring houses
or nearby trees and shrubs. The appropriate wind shielding correction
factor, "S" (from Table 2), is 0.9.
5. The house is new, and presumably wellbuilt. The appropriate leakiness
factor, "L" (from Table 3), is 1.4.
6. Calculate N:
N = 17 * 0.8 * 0.9 * 1.4
= 17
Calculate the average annual infiltration rate:
ACH = ACH50

17
= 14

17
= 0.82
The difference in this case (between dividing by 20 and 17) is not great
only 17%but it demonstrates how the building conditions and location can
affect the interpretation of pressurization tests.
Sherman compared his results to those reported by Persily. Sherman noted
that he obtained a correlation factor (N) of about 20 for a typical house
in the New Jersey area. Thus, Sherman's theoretically derived correlation
factor yields results similar to Persily's empirically derived correlation
factor.
The range of adjustment can be quite large. In extreme cases, the
correlation factor, N, can be as small as 6, and as large as 40. In other
words, the ACH50/20 rule of thumb could overestimate infiltration by a
factor of two, or underestimate it by a factor of about three.
This formula is still only a theory; it has not been validated with field
measurements. Moreover, there is considerable controversy regarding the
physical interpretation of the climate factor. For example, the formula
yields a yearround average infiltration rate, rather than just for the
heating season. Such a result is useful for houses with both space heating
and cooling, but it may be misleading for some areas.