600R95180
GEOMET Report Number IE-2603
March 9, 1995
ESTIMATION OF DISTRIBUTIONS FOR
RESIDENTIAL AIR EXCHANGE RATES
FINAL REPORT
prepared for
Patrick Kennedy
Exposure Assessment Branch
Economics, Exposure and Technology Division
Office of Pollution Prevention and Toxics
'U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D,C. 20460
under
EPA Contract Number 68-D9-0166
Task Number 3-19, Subtask 4
and
EPA Contract Number 68-D3-0013
Task Numbers 1-11 and 2-44
prepared by
t
GEOMET.Technologies, Inc.
20251 Century Boulevard
Germantdwn,- Maryland 20874
-------
DISCLAIMER
This document has been reviewed.and approved for publication
by the Office of Pollution Prevention and Toxics, U.S.
Environmental Protection Agency. The use of trade names or
commercial products does not constitute Agency endorsement or
recommendation for use...
11
-------
TABLE OF CONTENTS
Page
1. INTRODUCTION AND SUMMARY . ' 1
2. AIR EXCHANGE MEASUREMENTS 3
2.1. General Considerations, 3
2.2. Perfluorocarbon Tracers 4
3. DATA RESOURCES 9
3.1. Database of-PFT Ventilation Measurements -9
3.2. Extracts and Supplements for the , 11
PFT Database
4. ANALYSIS AND RESULTS . . ' ' 13
4.1. Data Exclusions and Supplements 15
4.2. Assignment of Weights 15
4.3. Summary Statistics / .17
5. REFERENCES . . '25
APPENDIX
Relationship Between Interzonal Airflows
and House Volume and Air Exchange
111
-------
LIST OF FIGURES
Page
1 Airflows for multiple-zone systems 6
2 Regions defined by the U.S. Census Bureau 13
3 Frequency .Distribution of Estimated Residential 2.0
Air Exchange Rates for All Regions Combined
VI
-------
ACKNOWLEDGEMENTS
This report was prepared by GEOMET Technologies, Inc.,
Germantown, Maryland, for the EPA Office of Pollution Prevention
and Toxics, Exposure Assessment Branch (EAB) , under EPA Contract
No. 68-D3-0013 (Task 2-44) with Versar, Inc., Springfield,
Virginia. The EPA-EAB Task Manager was Patrick Kennedy; his
support and guidance are gratefully acknowledged.
In addition to the authors of this report, .a number of
Versar/GEOMET personnel have contributed to this task over the
period of performance. These individuals are shown below:
Program Management - Gayaneh Contos, Versar
Task Management - Leo G. Schweer, Versar
Markeis, GEOMET
Editing - Kristeen Hinkley, GEOMET
Word Processing - Joy Warren, GEOMET
Glenda Heredia, GEOMET
VII
-------
1. INTRODUCTION AND SUMMARY
This report describes the information sources, analysis
methods and estimates of national and regional distributions for '
annual average air exchange rates measured in United States
residences. All the measurement results derive from techniques
involving constant release and time-integrated sampling of a
family of compounds known as perfluorocarbon tracers (PFTs).
This technique, developed at Brookhaven National Laboratory
(BNL), is described in Section 2 of this report. Section 2 also
provides a general discussion of air exchange, including
governing factors and alternative measurement methods.
Components of the PFT database used for this analysis,
although deriving from more than one field study, are unified by
common reliance on the BNL measurement protocol. Further, all
~\ a VN/-N va *- r^->~T.r T.T^-v-V -n -«^ *-3 v*-v- ? T*I -> *".*. ^~*~^ d^d"1""Cii3 'jLjrdi-i^inM C-L-Lk^J.
checking) was accomplished in a single laboratory. Prior to the
availability of the PFT database, relatively little work had been
directed toward estimating a national distribution for air
exchange rates, primarily due to the lack of national-scale
studies featuring unified protocols. The review chapter on
infiltration in the.current edition of the ASHRAE Handbook of
Fundamentals (ASHRAE 1993), for example, briefly summarizes a
number of air exchange studies that involved as many as 300
homes. Of these, only two relatively early studies (Grimsrud et
al. 1982, Grot and Clark 1979) are cited as giving a national
perspective, and these studies were restricted in terms of the
type of housing.
Since 1982, the PFT technique has been used to measure
airflow rates in more than 4,000 occupied residences in
accordance with BNL protocols. Under EPA sponsorship, these
measurement results were compiled by BNL into data files that
were used by Versar (1990) in developing a database and user
interface to allow various researchers to access the data.
Pertinent features of the database and interface are outlined in
Section 3 of this report. The structures represented in the
database do not constitute a random sample of residences across
the United States. Rather, they represent a compilation of
residences visited in the course of about 100. separate field-
research projects by.various organizations, some of which
involved random sampling and some of which involved judgmental or
fortuitous sampling. Certain areas of the country (e.g.,
Southern California) are heavily represented in the database used
for .analyses, whereas other areas have little or no
representation. Similarly, some seasons are more heavily
represented than others.
The geographic imbalance in the PFT database was partly
compensated by seeking additional PFT measurement results (e.g.,
from recently completed studies) for areas .with limited
-------
representation. Further compensation was obtained by applying
weighting factors in the analysis. These factors were developed
in such a way that state-specific results would contribute to
resulting estimates in proportion to their respective numbers of
occupied housing units, as determined from the 1990 census of
population and housing. The weighting factors, as well as
criteria for excluding certain cases from the analysis, are
presented in Section 4 together with the results of the analysis,
The results indicate that a value of 0.18 air changes per
hour (ACH) would be appropriate as a conservative number when
modeling inhalation exposure, and that a value of 0.45 ACH would
be appropriate when a typical air exchange rate is desired for
modeling inhalation exposure. These values correspond to the
10th and 50th percentile, respectively, for the estimated
national distribution of annual average air exchange rates.
-------
2. AIR EXCHANGE MEASUREMENTS
.2.1.. General Considerations
Air exchange, the balanced flow of air into and out of a
building, is an important determinant of indoor air quality. For
pollutants of indoor origin, air exchange can serve to dilute
indoor levels. Pollutants of outdoor origin, however, can also
be brought indoors. Air exchange is comprised of three
processes:
Infiltration - airflows through random cracks,
interstices and other unintentional openings in the
building envelope;
Nacurai ventilation - airflows through doors, operable
windbws, and other designed openings in the building
envelope; and
Forced ventilation - controlled airflows driven by
mechanical systems.
While natural ventilation and forced ventilation contribute
at times in homes, infiltration is the dominant mechanism for
residential air exchange. Infiltration is the inadvertent air
leakage of an otherwise sealed building. The use of natural
ventilation is seasonal, and forced ventilation systems
(excepting small local exhaust fans for spot ventilation in
kitchens and baths) are more often found in larger, non-
residential buildings.
Basic principles relating, to the physics, modeling and
measurements of air exchange have been summarized elsewhere
(ASHRAE 1993; Hutcheon and Handegord 1989; Sherman 1990).
Infiltration airflows are driven by pressure differences created
by wind action and indoor-outdoor temperature differences. The
driving pressure, in turn, acts upon the leakage sites
distributed over the building surface. Two basic avenues of'
measurement have evolved for air exchange: (1) pressurization
testing to measure the effective leakage area of the building
coupled with modeling to calculate air exchange, and (2) tracer-
gas techniques to monitor the dilution effects of air exchange.
Pressurization testing measures the effective leakage area
and thus provides information on a more or less constant property
of the building. While calculation methods exist to relate
effective leakage area to air exchange (see ASHRAE 1993), results
are predicated on empirical assumptions (e.g., physical
distribution of leakage area on the building surface) and
meteorological considerations.
-------
Tracer-gas techniques, on the other hand, provide a more
direct measure of air exchange. Tracer-gas techniques are
generally based on mass-balance considerations with explicit
assumptions that (1) well-mixed conditions prevail, and (2)
outdoor concentrations are negligible, and take the form:
dC _ St_ Q_c
dt ~ V V Ct
where
Ct = tracer-gas concentration at time t (e.g., mg/m3)
Sc = tracer-gas release rate at time t (e.g., mg/h)
V = volume of tested airspace (e.g., m3)
Qt = exiting airflow at time t (e.g., m3/h)
The ratio of exiting airflow (Q, m3/h) to indoor volume
(V, m3) gives the air exchange rate (I, h"1) . Three basic tracer-
gas strategies have evolved to measure air exchange :
(1) concentration decay, (2). constant concentration, and
(3) constant injection. The concentration-decay method involves
releasing a small amount of tracer gas, mixing it into the full
air space and monitoring the change in concentration with time.
From equation 1, the concentration data are related to the air
exchange rate by the analytical solution:
C2 = Cx e -1 (t>-t<), or I = Ln
-------
tracer technique, a kit was developed by researchers at
Brookhaven National Laboratory (BNL). The field components--
miniature perfluorocarbon tracer (PFT) permeation sources and
passive samplers--are inexpensive and reusable, require minimum
training of field personnel, and can provide integrated
measurements over days or months.
' 0
The PFT technique developed by BNL is an offshoot of the
constant-inject ion,, steady-state method of measuring air
infiltration that has been in use for many years. In that
scheme, a tracer is emitted,into a building at a constant known
rate and its concentration is allowed to reach a steady-state
level. This value is measured and then converted to a flow of
outdoor air into the building through the use of a single-zone,
mass-balance model.
BNL has also extended this method to address multizone
-czGurcnr.anto by deploying a dirrerent PFT in each zone.
Measurement of the concentrations of each tracer in all zones of
the building permits the calculation of the infiltration and
exfiltratien of outdoor air for each zone of the building as well
as the flows between zones. The system uses miniature permeation
tubes as tracer emitters and passive samplers termed Capillary
Atmospheric Tracer Samplers (CATS) to collect the perfluorocarbon
tracers. The passive samplers are returned to the laboratory for
analysis where the tracers are separated by gas chromatography
and quantified by an electron capture detector (GC/ECD). A
detailed description of this technique has been presented
elsewhere (Dietz et al. 1986).
As shown'in Figure 1, a single-zone building, because it is
flow-coupled only to the outdoors, has two flows. A two-zone
building, on the other hand, has each zone coupled to the
outdoors as well as to each other, giving a total of six flows.
Generalizing, a building composed of N well-mixed zones will have
(N-l) interzonal airflows for each zone plus an indoor/outdoor
flow pair for each zone, giving a total of N«(N-1) + 2N, or
N»(N+1) flows. The airflows are computed by inserting the
measured tracer concentrations and the known tracer emission
rates into the system of mass-balance and flow-balance equations
for each zone of the building.
Although the PFT methodology is relatively simple to
implement, it is subject to errors and uncertainties. The
general performance of the sampling and analytical aspects of the
system are quite good. That is, sample collection by the CATs
allied to GC/ECD analysis will measure the correct time-weighted
average tracer, concentration to within a few percent (Dietz et
al. 1986). Nonetheless, significant errors can arise when
conditions in the measurement scene greatly deviate from
idealizations calling for constant, well-mixed conditions.
Principal concerns focus on the effects of naturally varying air
exchange and the effects of temperature in the permeation source.
-------
Single-Zone
System
Two-Zone
System *
*.
Three-Zone
System
-V
N-Zone System Requires (N + 1) N Airflows
Figure 1. Airflows for multiple-zone systems,
-------
Sherman (1989) carried out an error analysis of the PFT
methodology using mathematical models combined with typical
weather data to calculate how an ideal sampling system would
perform in a time-varying environment. He found that for simple
single-story (ranch) and two-story plus basement (colonial)
layouts, seasonal measurements would underpredict seasonal
average air'exchange by 20 to 30 percent. Underprediction can
occur because the PFT methodology is measuring the effective
ventilation (the product of ventilation efficiency and air
exchange), and the temporal efficiency will generally be less
than Unity over averaging periods of this length. Sherman also
noted, however, that while the bias could have an impact on »
determining air exchange (absent knowledge of ventilation
efficiency) for calculating energy loads, the effective air
exchange term is directly relevant to determining average indoor
concentrations resulting from constant sources.
Sherman (1989) also noted that while multiple tracers
improve estimates of average effective ventilation in multiple-
zone buildings, the uncertainty associated with a given
interzonal airflow is 10 to 20 percent of the zone total,
reducing the quantitative utility of smaller estimated airflows.
As with the effective air exchange concept, however, the airflow
estimates correspond to effective interzonal transport and are
directly relevant to determining average indoor concentrations in
one zone resulting from a constant source in another.
Leaderer et al. (1985) conducted a series of experiments in
a room-sized environmental chamber to evaluate the practical
impacts of varying air exchange and the temperature response of
the permeation sources. The negative bias anticipated in the
measured (effective) versus actual air exchange as conditions
varied diurnally between 0.4 and 1.5 ACH was evident but minor
(3 to 6 percent), most likely due to the mechanical mixing in the
chamber and the relatively short'integration time (72 h).
Similarly, cycling temperature diurnally over an 8 C° range
(holding air exchange steady at 0.6 ACH) would cause
concentrations changes of about 20 percent as emissions
fluctuated. The investigators found; however, that using a time-
weighted average temperature to define the emission rate reduced
the temperature bias to essentially zero.
The PFT measurement system is also subject to potential
complications related to choices made when deploying the tracers.
If a field researcher deploying tracers unknowingly treats a
well-mixed zone as two separate zones, then the computational
solution for the ventilation flows becomes very unstable. In the
mathematical literature, the system is termed ill-conditioned.
Researchers at BNL developed a means to reliably detect and
accommodate such unstable situations using the matrix of tracer
concentrations to compute a condition number (D'Ottavio et al.
1988). Through the use of the condition number, large flow
errors for homes with two or more well-mixed zones can be
-------
errors for homes with two or more well-mixed zones can be
eliminated by combining .the well-mixed zones into a single zone
prior to the computations. As described later,- a number of
measurement results were encountered in this analysis for which
the automatic zone-reduction procedure had been applied at BNL.
-------
3. DATA RESOURCES
Most of the measurement results used for this analysis were
extracted from a database of PFT ventilation measurements
developed by Versar (1990) based on data files received from BNL
under a contract with the USEPA Office of Toxic Substances.
Selected features of this database are described in Section 3.1.
Information extracted from the database was supplemented with PFT
measurement results from three additional field studies. The
need for and nature of these supplements are discussed in
Section 3.2.
i
3.1. Database of PFT Ventilation Measurements
The PFT database consists of six distinct files:
{j.; me Project file, which' contains information about the
institution that performed the measurements and the
general location of the houses associated with the
project;
(2) The House file, which contains information associated
with the entire house such as the house setup code, the
start and stop dates for the measurement, and the
overall air exchange rate for each house;
(3) The Zone 1 file, which contains information unique to
the first zone in the measured building, including the
zone volume, the tracer used, and the infiltration and
exfiltration flow rates for the zone;
(4) The Zone 2 file, which contains information unique to
the second zone, if any;
(5) The Zone 3 file, which contains information unique to
the third zone, if any; and
(6) The Zone 4 file, which contains information unique to
the fourth zone, if any.
Many houses are associated with a particular project record
(i.e., there are many-to-one relationships between the House file
and the Project file). Each house is associated with one Zone 1
record since each house must have at least one zone in which PFT
measurements were taken. Therefore, a one-to-one relationship
exists between the House file and the Zone 1 file. In addition,
each house may or may not be associated with a record from the
Zone 2, Zone 3, or Zone 4 files depending on how many zones the
house was divided into for the ventilation measurement. For
example, a house in which two measurement zones were defined by
-------
the researchers would appear in the House file and in the Zone 1
and Zone 2 files, but this house would not appear in the Zone 3
or Zone 4 files.
Each house calculation can be identified by a unique field
that exists in the House file and all Its associated zone files.
This field is labeled HOFILE in the House file, Z1FILE in the
Zone 1 file, Z2FILE in the Zone 2, Z3FILE in the Zone 3 file, and
Z4FILE in the Zone 4 file. The House and Zone files can always
be related through this unique field. Further details on the
specific contents of these files are provided in a separate
report (Versar 1990).
Some houses appear in the database more than once for a
particular set of measurements because zone-reduction techniques
were applied by BNL. In certain cases where researchers
deploying PFTs unknowingly divided a well-mixed zone into two'
separate zones, relatively large errors were associated with the
airflow rates computed by BNL. The.large flow errors for homes
with two o"r more well-mixed zones were eliminated by combining
the well-mixed zones into a single zone prior to performing
computations on airflow rates.
In the quality-control review of the database, Versar and
GEOMET developed six "flag" fields to identify cases where zone-
reduction procedures were applied as well as other cases that
users might want to avoid. The names and explanations of these
six flags are as follows: '
(1) Zone Reduction Flag - A code of "1" indicates cases
where zone-reduction procedures were applied (such
houses are also represented in the database in reduced
form); a code of "0" indicates cases where zone-
reduction procedures were never applied as well as the
final reduced form for cases were zone-reduction
procedures were applied,-
\ TT c ^i = g . »i" indicates homc& uui_toide uhe united
States; "0" indicates homes within the United States;
(3) Zone Type Flag - "1" indicates cases where PFT
measurements were taken in unconditioned spaces such as
"attics, garages, or crawl spaces; although airflows to
and from such zones are of interest, the overall air
exchange rate for the residence can be misleading (and
usually will be abnormally large) because it
incorporates the air exchange between unconditioned
zones and outdoors;
10
-------
(4) Sums Flag - "1" indicates cases where the sum of
airflow rates into and out of each zone do not agree
within one percent;
(5) Negative Flows Flag - "1" indicates cases where one or
more negative airflows has been computed for a house;
although physically impossible, negative airflow rates
can result from calculations because of measurement
uncertainties; and
(6) Flow Error Flag - "1" indicates a case for which one or
more zones has an error associated with an estimated
airflow rate that is as large as, or larger than, the
flow itself; thus, the estimated airflow rate cannot be
confidently distinguished from zero.
Two additional sets of important information were included
among the files provided by BNL. First, a description of the
house structure was included in a specially designed house setup
code. This code indicates where each house zone is located,
whether the house has a fireplace, and whether the house has a
basement, crawl space, attic, or garage as part of the measured
space. Such information is useful in understanding (1) how the
zoning of a house was conceptualized by the field researchers and
(2) cases where unconditioned zones, which may have atypically
high ventilation rates, were included in the overall ventilation
measurement for the house.
The second major addition was information about the projects
for which the ventilation measurements were made. On average,.
each project comprised approximately 50 house measurements. A
group of ventilation measurements was associated with a
particular project name if all the measurements were conducted by
the same institution and if all the homes were located in the
same geographic area (i.e., within about 100 miles of each
other). Each project has a unique project code, a full
description of the location of the homes, and a short description
of the project. In addition, information is provided about the
institution that made the measurements, such as the names, phone
numbers and full addresses of contact people at the institution.
3.2. Extracts and Supplements for the PFT Database
The software developed by Versar for accessing the PFT
database provides three essential types of functions:
(1) retrieving and sorting data; (2) viewing data files; and
(3) creating reports and output files.
The function for creating reports and output files is
relatively powerful. User-selected reports can be directed to a
printer or disk file, enabling the user to examine information
11
-------
for the chosen houses in greater detail outside the database.
The user can also request that certain information about the
chosen houses be written to an ASCII file. The ASCII file is
written in a relatively straightforward format (i.e., single
record for each house consisting of fields in fixed locations
with a blank delimiter between fields). This format is
compatible with a number of commonly used software programs
(e.g., Lotus, dBase, SPSS/PC) and programming languages,
providing the user with the capability to .further manipulate or
.analyze the data outside the database. This feature is important
because the PFT database and software have no built-in capability
for statistical analysis. ;
The option to create an ASCII file was chosen to enable
analysis outside the PFT database. This option allows the user
to create an ASCII file either (1) from the House, Zone 1 and
Zone 2 files or (2) from the Zone 3 and Zone 4 files. Because
the focus of the analysis was on whole-house air exchange rates,
as opposed,, to airflows among zones within residences, the
selection was made that created a file from the House, Zone 1 and
Zone 2 files. The key information in this file needed for data
analysis included the state in which measurements were taken, the
project code, the measured whole-house air exchange rate, and the
date when each measurement was initiated. Prior to extracting
this information (together with additional information included '
in the file that was not needed for analysis), certain cases were
excluded. The nature of and reasons for these exclusions are
discussed in Section 4.
The primary intent of the analysis was to provide an
estimate of the national distribution of residential air exchange
rates. Preliminary analysis of the data indicated that (1)
measurements from the Los Angeles area (Wallace 1987, Wilson et
al. 1986) constituted a large fraction (more than one-third) of
the database entries and (2) there were few measurements from
states in the north-central and southern regions of the country
(as defined by the U.S. Census Bureau; see Figure 2). The
measurements from the Los Angeles area were supplemented with PFT
icoulLb f-LGiu Lvno icx-ciii- lield studies for the California energy
Commission (Berkeley Solar Group 1990 and ADM Associates 1990)
that were conducted primarily in other areas of the state.
Measurements for the southern region were supplemented with PFT
results from a 1991 study in Florida (GEOMET 1992). Potential
supplemental sources of information for the north-central region
of the country also were explored, but none were found.
i
The supplemental data were appended to the ASCII file
extracted from the PFT database, with critical information such
as state, air exchange rate, and measurement date entered in
appropriate field locations. Data analysis, described in the
next section, was applied to the supplemented ASCII file using
SPSS/PC software.
12
-------
WEST
NORTH CENTRAL
NORTHEAST
HAWAII
Figure 2. Regions defined by the U.S. Census Bureau.
13
-------
Other researchers* analyzing the PFT database discovered
some errors in the data from the SoCal field project (Wilson et
al. 1986). These errors, which pertained to the house volume,
zone-specific volumes and air exchange rate for some of the
houses, were found to be traceable to the original data files
provided to Versar by BNL. Corresponding data for these fields
were obtained from the original SoCal field researchers and added
to the PFT database. The air exchange rate was recalcu? ated by
dividing the house infiltration airflow rate from the PFT
database by the house volume provided by the researchers. In
cases where the recalculated air exchange rate differed from the
one in the original database by more than 10 percent, a flag was
generated and the air exchange rate provided by the original
researchers was used instead. Based on a series of contacts with
original researchers for other field projects who were able to
locate their own copies of PFT measurement results, it was
determined that any errors in the PFT database appear to be
confined' to the SoCal project.
Muhilan Pandian, University of Nevada Las Vegas,
letter communication dated August 1, 1994.
14
-------
4. ANALYSIS AND RESULTS
4.1 Data Exclusions and Supplements
The analysis procedure involved (1) exclusion of certain
cases that could have biased the results and (2) assigning
weights to the results from each state to compensate for
geographic imbalance in the locations where PFT measurements were
taken. The original PFT database compiled by Versar from BNL
data files contains 4,590 measurement results. However, as noted
in Section 3, some homes are represented more than once (i.e.,
essentially redundant results for whole-house air exchange rates)
because zone-reduction procedures were applied; in such
instances, only the reduced case was used (i.e., the case with
the smallest apparent measurement error). Residences outside the
U^iLcJl .jwctLco WC.LC: excluded, as were nomes with unconditioned
spaces (which could lead to artificially high air exchange
rates). Also excluded were small field studies involving
repeated measurements in research ho'uses, based on descriptions
provided in the Project file.
After the above exclusions, 3,174 cases remained for
-analysis. EPA-sponsored TEAM studies involving a random cross-
section of houses in California, Maryland and New Jersey involved
repeated measurements in a subset of the houses, typically for
two sequential 12-hour intervals (daytime and nighttime). The
house code assigned by the original researchers* enabled
identification of these sequential measurements, which were
averaged to provide a single air exchange rate for each house.
Following this averaging process, which eliminated 285 repeat
measurements, there were 2,889 cases for analysis. Supplemental
measurements from California and Florida, described in the
previous section, increased the number of cases to 2,976.
These cases were first examined in terms of state-specific
frequency distributions. A few negative values that were found
were excluded from further analysis (such values are physically
impossible). Also excluded were several apparent outliers--three
for Texas and one each for California and Maryland. In each
case, the excluded values were at least seven times greater than
the next highest value for the state. After excluding these
outliers, 2,971 cases remained for analysis.
4.2 Assignment of Weights
The remaining data were examined by state and by the four
geographic regions defined by the U.S. Census Bureau (see
Figure 2). As shown in Table 1, measurements taken in California
accounted for nearly half of all measurements in the database and
Lance Wallace, U.S. Environmental Protection Agency,
Warrenton, VA, personal communication.
15
-------
Table 1. Original and Weighted Numbers of PFT Measurements,
by Region and State
Region/State
West Region
Arizona
California-Los Angeles
California-Other
Colorado
Idaho
Montana
Oregon
Washington
Region Total
;North Central -Region,
Minnesota
Wisconsin
Region Total
Northeast Region
Connecticut
New Jersey
New York
Region Total
South Region
Florida
Maryland
Texas
Region Total
Total , All Regions
Number of Valid
PFT Measurements
25
1,388
69
9
78
51
212
235
2,067
'
63
57
120
25
9
527
561
t>
18 .
163
42
223
2,971
Weight Applied
0.680
0.046
0.928
1.778
0.051
0.078
0.066
0.098
1.825
2.246
0.960
6.000
0.241
'.'''.".'.,
7.611
0.288
3.857
Resultant
Number
of Cases
l
17
64
64
16
4
' ' 4
14
23
206
115
128
243
24
54
'. 127
205
137
47
162
346
1,000
16
-------
two-thirds of the measurements in the west region. To prevent
this preponderance of results for California from excessively
influencing summary statistics, weights were assigned to each
state's results so that the resultant number of cases would
represent each state in proportion to its share of occupied
housing units, as determined from the 1990 census of population
and housing (U.S. Bureau of the Census, 1991). Further, for
California, the weights for the Los Angeles area and remainder of
the state (which are nearly equal in housing share) were set so
that each of these areas would contribute equally to the weighted
results. The weights for each,state also reflected each region's
share of occupied housing units. The weights were constructed in
such a way that the resultant number of cases, when summed across
regions, would total exactly 1,000 (see Table 1).
4.3 Summary Statistics
Summary statistics for estimated annual average air exchange
rates, based on weighting of cases as described above, are given
in Table 2 for each of the four geographic regions and for all
regions combined. The air exchange rates are in units of air
changes per hour (ACH). Up through the 65th percentile, the
distributions are fairly similar across geographic regions (i.e.,
the values for each region are within 0.1 ACH of the value for
all regions combined). The primary differences across regions
are lower-than-average values for the north-central region and
higher-than-average values for the northeast region, as reflected
by both the arithmetic and geometric means.
For purposes of modeling inhalation exposure in residential
settings, the 10th percentile value of 0.18 ACH for all regions
combined seems appropriate when a conservative air exchange rate
is desired. None of the region-specific values at this
percentile deviates markedly from 0.18. When a typical air
exchange rate is desired for modeling, the median (50th
percentile) or geometric mean is a preferred measure of central
tendency because neither is excessively,influenced by extreme
values at the upper tail of the distribution. The median value
.for all regions combined is 0.45 ACH and the geometric mean is
0.46 ACH. Thus, 0.45 ACH seems appropriate as a typical air
exchange rate.
In applying conservative or typical values of air exchange
rates for modeling purposes, it is important to keep in mind the
limitations of the underlying data base. Although the estimates
are based on thousands of measurements, the residences
represented in the data base are not a random sample of the
United States housing stock. The houses are not geographically
balanced, but efforts have been made to compensate for this
imbalance. Further, as discussed later, the results are not
equally balanced among different times of the year, and air
exchange rates often vary seasonally because of changes in
17
-------
Table 2. Summary Statistics* for Estimated 'Annual Average
Air Exchange Rates (in ACH) by Region
Statistic
Unweighted
No. of Cases
Weighted
No. of Cases
Arithmetic
Mean
Arithmetic
Standard
Deviation
Geometric
Mean
Geometric
Standard
Deviation
West Region
2067
206
0.66
0.87
0.47
2.11
North
Central
Region
120
243
0.57
0.63
0.39
2.36
Northeast
Region
561
205
0.71
'0.60
0.54
2.14
South
Region
223
346
0.61
0.51
0.46
2.28
All
Regions
2,971
1,000
0.63
0.65
0.46
2.25
Percentiles of Distribution
5th
10th
15th
20th
25th
30th
35th
40th
45th
50th
55th '
OULil
65th
70th
75th
80th
85th
90th '
95th .
Maximum
0.16
0.20
0.23
0 . 27
0.30
0.30
0.33
0.37
0.40
0.43
0.48
U . J^.
0.60
0.68
0.76
0.83
0.96
1.25
1.96
23.32 .
0.12
0.16
0.18
0.20
0.22
0.23
0.26
0.29
0.32
0.35
0.38
^ . -1 _
0.51
0.55
0.62
0.75
1.00
1.49
1.92
4.52
0.19
0.23
0.27
0.32
0.32
0.36
0.39
0.42
0.48
0.49
0.53
: . : :
0.72
0.91
0.99
1.19
1.26
1.33
1.87
5.49
0.15
0.16
0.19
0.21
0.24
0.31
0.35
0.41
0.48
0.49
0.56
: - r
0.65
0.70
0.73
0.86
1.06
1.21
1.58
3.44
0.15
0.18
0.20
0.23
0.26
0.30
0.33
0.36
0.40
0.45
0.49
n r A
0.60
0.66
0.76
0.90
1.09
1.26
1.74
23.32
* All numbers reflect weighting by state as shown in Table 1,
18
-------
driving forces such as indoor-outdoor temperature differences.
Despite such limitations, the estimates in Table 2 are believed
to represent the best available information on the distribution
of air exchange rates across United States residences throughout
the year.
Components of the PFT database used for this analysis,
although deriving from more than one field study, are unified by
common, reliance on the BNL measurement protocol. Further, all
laboratory work and primary data analysis (including error
checking) was accomplished in a single laboratory. Prior to the
availability of PFT database, relatively little work had been
directed toward estimating a national distribution for air
exchange rates, primarily due to the lack of national-scale
studies featuring unified protocols. The review chapter on
infiltration in the current edition of the ASHRAE Handbook r^f
r^iJaiiiciii-calfi. v'/iortK/ir, ly^j; , ror example, briefly summarizes a
number of air exchange studies that involved as many as 300
homes. Of these, only two relatively early studies (Grimsrud et
al. 1982, Grot and Clark 1979) are cited as giving a national
perspective", and these .studies were restricted in terms of the
type of housing.
The estimated annual average distribution for all regions
combined is shown in graphical form, as a histogram with
intervals of 0.1 ACH, in Figure 3. The shape of the histogram
generally is consistent with that from a lognormal distribution.
Nearly 50 percent of the residential air exchange rates are in
the relatively narrow range of 0.1 or less to 0.4 ACH, but the
tail of the distribution is quite extended and includes about
3 percent of residential air exchange rates that exceed 2.0 ACH.
Summary statistics for estimated seasonal average air
exchange rates are given in Table 3. Seasons were defined by
grouping December, January, and February measurements-for winter;
March, April and May for spring; and so on. Caution should be
exercised in interpreting or applying the seasonal statistics
because (1) the seasons are not equally represented in terms of
unweighted or weighted numbers of cases, and (2) attempts to
compensate for geographic imbalance were not applied on a season-
specific basis. The winter and spring measurements each account
for more than a third of the total observations whereas the fall
measurements account for less than 10 percent. Limitations
related to seasonal imbalance apply equally to the annual
estimates previously given in Table 2. Although, in concept,
efforts could have been made to adjust for both geographic and
seasonal imbalance in developing the annual estimates, such an
effort would have been severely hampered by the limited
information available for numerous combinations of geographic
location and time of year.
19
-------
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
AIR EXCHANGE INTERVAL, ACH
>2.0
Figure 3. Frequency Distribution of Estimated Residential
Air Exchange Rates for All Regions Combined.
20
-------
Table 3. Summary Statistics* for Estimated Seasonal
Average Air Exchange Rates
Statistic
Unweighted
No. of Cases
Weighted
No. of Cases
Arithmetic
Mean
Arithmetic
Standard
^ c: * _L. a v~ j. s-^ii
Geometric
Mean
Geometric
Standard
Deviation
Winter
1162
439
0.67
0.60
0.52
2.01
Spring
1160
387
0.65
0.65
0.45
2 . 41
Summer
507
121
0.57
0.88
0.34
' 2.48
Fall
142
53
0.38
0 . ^1
0.31
1.88
Percentiles of Distribution
10th
25th
50th
75th
90th
0.21^
0.34
0.51
0.86
1.26
0.19
0.27
0.43
0.76
1.50
0.15
0.16
0.26
' 0.63
1.45
0.14 -
0.22
0.30
0.48
0.74
*A11 numbers reflect weighting by state as shown.in Table 1.
21
-------
Seasonal and annual average air exchange rates are
summarized by region and by state in Table 4 (arithmetic mean and
standard deviation) and Table 5 (geometric mean and standard
deviation). The unweighted and weighted numbers of cases for
each area-season combination are also shown in Table 4. In most
cases the seasonal means are below one ACH. The major exception
is the Los Angeles area, for which both the arithmetic and the
geometric mean for summer are above one ACH (most likely due to
open-house configuration under moderate climate conditions). New
Jersey also has arithmetic/geometric means of one ACH or higher
for winter, but these statistics are based on a limited set of
nine homes.
As noted previously, the estimates provided in this section
are believed to represent the best available information on
distributions of air exchange rates but are limited by both
geographic and seasonal imbalances. Such limitations could
conceivably be overcome by conducting additional measurements in
areas and at times that have received limited coverage to date.
Given sufficient resources, a preferred alternative would be to
select- a nationally representative sample of residences and
conduct week-long measurements that are equally distributed
throughout the year in each region of the country.
In addition to the air exchange rate for the house,
estimates of air flows among zones within a house may be needed
to model inhalation exposure, because the release of airborne
chemicals is often confined -to one zone of the house. Given the
variety of situations represented in the PFT database, it is
difficult to choose typical or conservative values for interzonal
airflow rates. Consequently, an effort was made to model the
relationship between interzonal airflow rates and whole-house air
exchange rates. This analysis effort, described in the Appendix,
was conducted for two situations--bedroom area versus remainder
of the house and kitchen area versus remainder of the house.
22
-------
Table 4. Arithmetic Statistics for Estimated Seasonal and
Annual Average Air Exchange Rates, by Region and State
Region/State
West
Region
Arizona
California-
Los Angeles
California-
Other
Colorado
Idaho
Montana
[ . i
Oregon
Washington
North Central
Region
Minnesota
Wisconsin
Northeast
Region :
Connecticut
New Jersey
New York
South
Region
Florida
Maryland
Texas
Winter
0.51 ± 0.88*
(752,72)
,^^^_MB^B^ ^MB mMBMMi^^H
0.40 + 0.23
(11,7) %
0.77 + 1.72
(375,17) .
0.56 + 0.33
(16,15)
0.24 +0.07
(2,4)
0.50**
no. i ^
0.21 ± 0.22
(27,2)
a
0.49 ± 0.45
(128,8)
0.32 + 0.19
(169,17)
0.36 ± 0.24
(28,53)
0.27 ± 0.12
(23,42)
0.70 +0.28
(5,11)
0,76 :+. 0.50 :
(342,144) '
0.44 + 0.31
(14,13)
1.12 ± 0.46
(9,54)
0.55 ± 0.40
(319,77)
. 0::75:::;±::"d.:55 ;:
,140,170) : ;
0.43 + 0.14
(8,61)
0.60**
(4,1)
0.93 + 0.62
(28,108)
Spring
0.60 ± 0.56
(692,49)
^HM«^H««B^«MH^BH^^BMM^MM
0.49**
(1,1)
0.81 + 0.68
(551,25)
0.31 + 0.02
(3,3)
0.33 ± 0.22
(6,11)
^0.55 ± 0.66
f ~*. -, "> '
0.20**
(4,0)
0.51 ± 0.35
(51,3)
0.36 ± 0.23
(41,4)
0.66 ± 0.70
(87,181)
0.34 ± 0.32
(35,64)
0.83 + 0.79
(52,117)
0.61 ± 0.78
(210,;59)
1.25 + 1.63
.(11,11)
--
0.47 + 0.31
(199,48)
;;:o .;68 :;± o.:47 ;
.' (171,: 99). ;
.
0.59 + 0.44
, (157,45)
0.75 + 0.49
(14,54)
Summer
0.98 ± 1.14
(492,58)
*«BIWM^HI«WMI^M.^^_
0.36 + 0.20
(10,7)
1.62 + 1.62
(449,21)
0.68 + 0.50
(33,31)
--
.--
--
-
--
--
--
--
0.82**.
(5,1) :.
--
--
0.82**
(5,1)
0.19 ±:0.:05
v 01 0/61) :.
0.19 + 0.04
(8,61)
0.28**
(2,1)
--
Fall
0.47 ± 0.38
(131,27)
MBMM^HIBHMBIMMHBMHHM^^^
0.41 ± 0.36
(3,2)
0.85**
(13,1)
0.52 + 0.28
(17,16)
0.16 ± 0.00
U',2)
0.36**
119,1)
0.25 + 0.25
(20,2)
0.63 ± 0.42
(33,2)
0.39 ± 0.18
(25,2)
0.13 ± 0.03
<5,9)
0.13 + 0.03
(5,.9)
--
:0.23**
(4,:l)
'--
--
0.23**
(4,1)
,0.36 ± 0.12
, (2,15)
0.36 ± 0.12
(2,15)
--
.--
All Seasons
0.66 ± 0.87
(2067,206)
I^^^BMHMB>*«M^H^HIMI^HH»KIIBd
0.39 + 0.21
(25,17)
1.06 ± 1.39
(1388,64)
0.60 + 0.41
(69,64)
0.29 + 0.19
(9,16)
n 4Q j. n c-
(78~,4)
0.23 ± 0.17
(51,4)
0.52 ± 0.40
(212,14)
0.34 ± 0.19
(235,23)
0.57 ±0.63
(120,243)
0.30 ± 0.26
(63,115)
0.82 +0.76
(57,128)
0.71 ± 0.60
(561,205')
0.80 + 1.15
(25,24)
1.12 ± 0.46
(9,54)
0.52 ± 0.37
(527,127)
0.61 ± 0.51
(223,346)
0.31 ± 0.16
(18,137)
0.59 ± 0.44
(163,47)
0.87 ± 0.58
(42,162)
Arithmetic mean + standard deviation (unweighted and weighted number
cases); all numbers reflect weighting by state as shown in Table 1.
Based on one or fewer cases after weighting.
of
23
-------
Table 5. Geometric Statistics for Estimated Seasonal and
Annual Average Air Exchange Rates, by Region and State
Region/State .
West
Region
Arizona
California-
Los Angeles
California-.
Other
Colorado
Idaho
Montana
Oregon
Washington
North Central
Region
Minnesota
Wisconsin
Northeast :.
Region
Connecticut
New Jersey
New York
South '
Region :
Florida
Maryland
Texas , -
Winter
0.39 ;± .1.93*
0.35 ± 1.62
0.54 ± 2.10
0.49 ± 1.70
0.23 ± 1.37
0.31**
0.18 ± 2.04
0.41 ± 1.81
0.28 ± 1.75
0.30 ± 1.75 '..
0.25 ± 1.48
0.65 ± 1.49
.0.:59 ± :2.12. ':
0.37 ± 1.81
1.00 ± 1.71
0.45 .± 2.03
0.63 ±.1.77
04] 0.1 at
0.56**
0.80 ± 1.70
Spring
0,46 ± 2.00 ;
0.49**
0.65 ± 1.97
0.31 ± 1/07
.0.28 ± 1.63
0.41 ± 3.29
0.20**
0.46 ± 1.70
0.31 ± 1.82
0.44 ± 2.46
. 0.27 + 2.00
0.57 ± 2.43
.0:44 ±2:00 :;
0.67 ± 3.10
--
0.40 ± 1.82 .
0.49 ± 2.72
0.42 ± 2.88
0.55 ± 2.57
Summer
0.67 ± 2.31
0.31 ± 1.76
1.12 ±2.46
0.56 ± 1.88
--
--
--
--
--.-.. . '.
' --
.
0.64** : -.';
--
0.64**
.0.18 ± 1:.38
2 . ic x i . 11
0.05**
--
Fall
0.34 ±. 1.81
0.34 ± 2.26
0.45**
0.47 ± 1.60
0.16 ± 0.00
0.28**
0.21 ± 2.57
0.55 ± 2.05
0.36 ± 1.57
0.13 ± 1.25.
0.13 ± 1.25
--
0.22**
--
--
0.22**
.0.34 ± 1.43
u . 0-i ± J. . 4^
--
--
All Seasons
0.47 ± 2.11
0.34 ± 1.67
0.73 ± 2.27
0.50 ± 1.76
0.25 ± 1.58
0.34 ± 2.60
0.19 ± 1.84
0.44 ± 1.75
0.29 ± 1.71
0.39 ± 2.36
0.24 ± 1.82
0.58 ± 2.35
0.54 ± 2.14
0.48 ± 2.46
1.00 ± 1.71
0.43 ± 1.95
0.46 ± 2.2E
0.28 ± 1.62
0.41 ± 2 . 99
0.71 ± 2.05
Geometric mean ± standard deviation; all numbers reflect
weighting by state as shown' in Table 1.
Based on one or fewer cases after weighting. '
24
-------
5. REFERENCES
ADM Associates. 1990. Pilot Residential Air Exchange Survey.
Task 2; Pilot Infiltration Study. Indoor Air Quality Assessment
Project. ADM Associates, Inc., Sacramento, CA. Prepared for
California Energy Commission under Contract No. 400-88-020.
ASHRAE. 1993. Infiltration and Ventilation. In 1989 ASHRAE
Handbook of Fundamentals. American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Atlanta, pp. 23.1-
23.20.
Berkeley Solar Group. 1990. Occupancy Patterns and Energy
Consumption in New California Houses (1984-1988). Berkeley Solar
Group and Xenergy, Oakland, CA. Prepared for California Energy
Commission under Contract No. 400-87-015.
t
D'Ottavio, T.W., G.I. Senum, and R.N. Dietz. 1988. Error
Analysis Techniques for Perfluorocarbon Tracer Derived Multizone
Ventilation. Rates. Building and Environment 23:4, pp. 187-194.
Dietz, R.N., R.W. Goodrich, E.A. Cote, and R.F. Wieser. 1986.
Detailed Description and Performance of a Passive Perfluorocarbon
Tracer System for Building Ventilation and Air Exchange
Measurements. Measured Air Leakage of Buildings. ASTM STP 904,
H.R. Trechsel and P.L. Lagus, Eds., American Society for Testing
and Materials, Philadelphia, PA, pp. 203-264.
Dixon, W.J., and F.J. Massey, Jr. 1969. Introduction to
Statistical Analysis. McGraw-Hill, New York, NY.
GEOMET. 1992. Evaluation of Radon-Resistant Construction
Practices in New Homes in Florida. Report No. IE-2588, GEOMET
Technologies, Inc., Germantown, MD. DCA Contract No. 91 RD-41-
14-00-009.
\
Grimsrud, D.T., M.H. Sherman., ard R.C. Sonderegger. 1982.
Calculating Infiltration: Implications for a Construction
Quality Standard. Proceedings of the ASHRAE-DOE Conference on
the Thermal Performance of the Exterior Envelope of Buildings II,
Las. .Vegas, NV.
Grot, R.A. and R.E. Clark. 1979. Air Leakage Characteristics
and Weatherization Techniques for Low-income Housing.
Proceedings of the ASHRAE-DOE Conference on the Thermal1
Performance of .the Exterior Envelopes of Buildings, Orlando, FL.
Hutcheon, N.B. and G.O.P. Handegord. 1989. Building Science for
a Cold Climate. Construction Technology Centre Atlantic, New
Brunswick, Canada.
25
-------
Leaderer, B.P., L. Schaap, and R.N. Dietz. 1985. Evaluation of
Perfluorocarbon Tracer Technique for Determining Infiltration
Rates in Residences. Environmental Science & Technology. Vol.
19, No. 12, pp. 1225-1232.
Sherman, M.H. 1989. Analysis~.of Errors Associated with Passive
Ventilation Measurement Techniques. Building and Environment.
Vol. 24, No. 2, pp. 131-139. -
Sherman, M.H. (Ed.). 1990. Air Change Rate and Airtightness in
Buildings. ASTM STP 1067, American Society for Testing and
Materials, Philadelphia, PA.
i
U.S. Bureau of the Census. 1991. Statistical Abstract of the
United States: 1991. Washington, DC. Table No. 1284, p. 727.
Versar. 1990. Database of PFT Ventilation Measurements:
Description and User's Manual. Versar Inc., Springfield, VA.
Prepared for USEPA Office of Toxic Substances Under Contract No.
68-02-4254, Task No. 39.
Wallace, L.A. 1987. The Total Exposure Assessment Methodology
(TEAM) Study: Summary and Analysis (Volume 1). Report Number .
EPA/600/6-87/002a,< Washington, DC.
Wilson, A.L., S.D. Colome, P.E. Baker, and E.W. Becker. 1986.
Residential Indoor Air Quality Characterization Study of Nitrogen
Dioxide. Phase I. Volume 2; Final Report. Prepared for Southern
California Gas Company, Los Angeles, CA.
26
-------
Appendix
Relationship Between Interzonal Airflows
and House Volume and Air Exchange
-------
A.I. Introduction
Estimates of the house>volume and indoor-outdoor air
exchange rate may not be sufficient for modeling inhalation
exposure because the release of many airborne chemicals may be
confined to certain areas or zones within the house. In such
cases, the .airborne concentration within the zone of release
generally will be higher than in other areas of the house,
especially during the release period and until such time as the .
chemical has been completely mixed throughout the house.
The communication of air between one zone of the house and
the remainder is governed by a number of factors, including
(1) communication barriers such as walls, (2) door and window
openings that promote air movement to, from and within the house,
_. ,-; ^ f*3\ ^^^-^j-^-^.^^-^^^^-, ^ £ ^^^_ _, , .4~,i~ j -. ~ - '. ., ^ _. -, ._-v -_ -. ^ -. _i ~" _ _ ' .
.'._. L . '-,,, ..._ . _...,_.. .^ *_* ^ » _ <^ ^. iw* *_; o* N * * L* »j j- w j- w w *w± d J. j.
heating/cooling systems and portable or stationary fans. It is
hypothesized that the rate of air movement within a.house will
have some re-lationship to the indoor-outdoor air exchange rate,
.because the air exchange rate will tend to be higher (1) when
doors or windows are open, which will also promote air movement
within the house, or (2) when temperatures are colder outdoors,
which will place an increased demand on heating systems that move
air either mechanically or through convective heat transfer.
A. 2. Methods.
An empirical relationship between interzonal airflows and
the indoor-outdoor air exchange rate was developed using the PFT
database described in the main body of this report. Two
situations were examined: (1) bedrooms, for which communication
with the remainder of the house may be restricted because of
relatively small communication pathways (doors), and (2) the
kitchen, which generally has a more open communication path to
adjacent areas such as the living or dining room. The PFT
database was searched for cases where researchers labeled either
the kitchen or bedroom(s) as a separate zone, using the same
exclusion criteria as described in Section 4 of this report.
Based on these criteria, 408 bedroom cases and 360 kitchen cases
were located. The information extracted for these cases was the
house volume (m3) , the air exchange rate (h"1) and the airflow
rate (m3h"1) between the bedroom (s) or the kitchen and the
remainder of the house. .
This analysis could easily be confounded by different
volumes across houses or by inequalities in the airflow rates
into and out of the zone under study. Consequently, the inflows
and outflows for each house were averaged, and the averaged flow
was normalized by dividing it by the house volume:
A-l
-------
QN = (Q» + Qai)/2 (A_1}
V
where
QN is the normalized airflow between zone 1 (kitchen
or bedroom) and zone 2 (remainder of the house)
Q12 is the measured airflow rate from zone 1 to zone 2
Q21 is the measured airflow rate from zone 2 to zone 1
V is the volume of the house.
The normalized interzonal airflow rate, QN, has the same
units (h"1, or air changes per hour) as the air exchange rate.
The relationship of QN to the air exchange rate (I) was obtained
by linear regression analysis, yielding an equation of the-
following form:
. QN = a + bl ' (A-2)
where - .
a is the regression intercept (h"1)
b is the regression slope.
A.3. Results and Applications
For the bedroom case, the following relationship was
obtained from the regression analysis:
QN = 0.078 + 0.311 (R2 = 0.56) (A-3)
The relationship for the kitchen case was:
QN = 0.046 -I- 0.391 (R2 = 0.58) (A-4)
Thus, more than 50 percent of the variance across houses in the
normalized interzonal airflow rate was explained ,by each
regression equation.
Based on the relationships given above, characteristic
airflow rates can be postulated for two-zone situations
conceptualized as "bedroom versus remainder of the house" and
"kitchen versus remainder of the house." Consider, for example,
a house with a volume of 408 m3 (national average, as determined
from information reported from the Residential Energy. Consumption
Survey conducted by the U.S. Department of Energy*) and an air
exchange rate of 0.45 h'1 (national average, as listed in Section
USDOE. 1992. Housing Characteristics 1990. Report
No. DOE/EIA-0314(90), U.S. Department of Energy, Energy
Information Administration, Washington, DC.
A-2
-------
4 of this report) . From equation A-3, QN for the bedroom would
be 0.078 + 0.31 x 0.45, or 0.2175 h'1. Multiplication by the
house volume yields an interzonal airflow rate of 88.7 m3h"1. By
a similar process, the airflow rate between the kitchen and the
remainder of the house is calculated from equation'A-4 to be
90.4 m3}!'1.
One cautionary note is in order when using the relationships
described above. Some or many of the researchers contributing
measurements to the PFT database may have defined a zone as a
group of adjacent bedrooms, i rather than an individual bedroom.
If so, then the interzonal airflow rate for an individual bedroom
is likely to be lower than indicated by equation A-3. Similarly,
the living room, which like the kitchen has a fairly open
communication with the rest of the house but also has a larger
volume than the kitchen, might be expected to have a hiaher
race cnan indicated Cy equation A-4.
A-3
-------
6027M01 .
REPORT DOCUMENTATION
PAGE
4. Tide and Subtitle
1. REPORT NO.
Estimation of Distributions for Residential Air Exchange Rates
3. Recipient's Accession No.
5. Report Date
March 1995
6.
7. AuthorU)
M.D. Koontz, H.E. Rector
8. Performing Organization Rept. No.
GEOMET Report No. IE-2603
a. Performing Organization Name and Address
GEOMET Technologies, Inc.
20251 Century Boulevard
Germantown, MD 20874
10. Project/Task/Work Unit No.
11. Contract(C) or Grant(G) No.
(O 68-D9-0166
68-D3-0013
12. Sponsoring Organization Name and Address
United States Environmental Protection Agency
Office of Pollution Prevention and Toxics
Economics, Exposure and Technology Division
Washington. DC 20460
13. Type of Report & Period Covered
Final Report
14.
15. Supplementary Notes
The EPA Work Assignment Manager was Patrick Kennedy.
16. Abstract (Limit: 200 words)
This report describes the information sources, analysis methods and estimates of national and regional
distributions for annual average air exchange rates measured in United States residences. All the measurement
results derive from techniques involving constant release and time-integrated sampling of a family of
compounds known as perfluorocarbon tracers (PFTs).
The results indicate that a value of 0.18 air changes per hour (ACH) would be appropriate as a conservative number
when modeling inhalation exposure, and that a value of 0.45 ACH would be appropriate when a typical air exchange
rate is desired for modeling inhalation exposure. These values correspond to the 10th and 50th percentile,
respectively, for the estimated national distribution of annual average air exchange rates.
17. Document Analysis a. Descriptors
Exposure Assessment
Indoor Air Quality Monitoring
Indoor Air Quality Modeling
b. Identifiers/Open-Ended Terms
Air Exchange
Air Leakage
Tracer Gas
C. COSATI Reid/Group
IS.JWailability Statement
istribution Unlimited
19. Security Class (This Report)
Unclassified
20. Security Class (This Page)
Unclassified
21. No. of Pages
42
22. Price
(See ANSI-Z39.18)
See Instructions on Reverse
OPTIONAL FORM 272 (4-77)
(Formerly NTIS-35)
Department of Commerce
------- |