EPA/600/A-95/079
MODELING ENVIRONMENTAL TOBACCO SMOKE IN THE HOME USING
TRANSFER FUNCTIONS
Wayne R. Ott
U.S. Environmental Protection Agency
Atmospheric Research and Exposure Assessment Laboratory and
Department of Statistics, Stanford University
Stanford, CA 94305
Neil E. Klepeis
Information Systems and Sciences, Inc.
4220 South Maryland Parkway, Suite 311
Las Vegas, NV 89119
Paul Switzer
Department of Statistics
Stanford University
Stanford, CA 94305
Paper No. 95-WP84B.03 for presentation at the 88 th Annual Meeting of the
Air and Waste Management Association in San Antonio, TX, June 1995.
-------
ABSTRACT
This paper presents the theoretical and practical development of a multi-compartment indoor
air quality model designed for predicting pollutant concentrations from environmental tobacco
smoke (ETS) in the home. The model is developed using transfer functions for each
compartment, thereby obtaining analytical solutions that can be expressed mathematically and
do not require a computer. The input parameters to the model are the cigarette source
emission rate, smoking activity patterns, room volumes, compartmental air exchange rates,
and intercompartmental flow rates. Field experiments are conducted in an unoccupied home
using a cigar and cigarettes as sources to evaluate the performance of the model, and real-
time measurements are made in the home of carbon monoxide (CO), respirable suspended
particles (RSP), and polycyclic aromatic hydrocarbons (PAH). The time series predicted
from the equations by the model agree well with the concentration time series measured in
the rooms of the home. The transfer function approach can be applied to any home simply
by inspecting the floor plan and then writing the transfer functions by following simple rules.
TTie experimental data show that the door and window positions in each room exert
considerable influence on the pollutant concentrations observed in the home.
-------
INTRODUCTION
95-WP84B.03
Over a 24-hour period, people spend the greatest share of their time indoors at home. In
a statewide survey of activity patterns1, Californians over the age of 11 spent 63% of their time
at home, 25% in other indoor locations (offices, stores, etc.), 7% in enclosed vehicles (buses,
vans, automobiles, etc.), and only 5% outdoors. Many indoor sources of air pollution -
smoking, cooking, consumer products, gas appliances, building materials — are found in homes.
Despite the importance of the home microenvironment in contributing to one's total exposure to
environmental pollutants, relatively few indoor air quality models have been applied to the home
microenvironment and validated with experimental data from real homes.
This paper explores a multi-compartment indoor mass balance model that is adapted to a
small, 2-bedroom home to predict the concentration time series from environmental tobacco
smoke in various rooms. The home was temporarily unoccupied, allowing a variety of
experiments to be conducted with different combinations of source locations, monitoring
locations, door, and window positions. Transfer functions were used to predict the relationships
among the time series of concentrations in different rooms, and the model's performance was
evaluated using experimental time series data on carbon monoxide (CO), respirable suspended
particles (RSP), and polycyclic aromatic hydrocarbons (PAH).
DEVELOPMENT OF AN INDOOR TIME SERIES MODEL
McKone2 developed a three-compartment model to compute the 24-hour concentration
history of Volatile Organic Compounds (VOCs) in the shower, bathroom, and remaining
household volumes from tap water use. Wilks et al.3 used an indoor model implemented on a
personal computer to show that daily inhalation exposure for a person with a contaminated water
supply could exceed the person's daily ingestion exposure from the same tap water. Axley and
Lorenzetti4 used a multi-compartment indoor air quality modeling system based on commercially
available software. Sparks et. al5 developed a multi-compartment computer model for a home and
verified the model for VOC sources such as wood stain, varnish, and floor wax. Our approach is
to develop the model for environmental tobacco smoke (ETS) in a home using a transfer function
approach that yields analytical solutions and does not require a computer.
The mass balance model accounts for all the pollutant mass that either is emitted,
deposited, or mixed into the interior air of an enclosed compartment (for example, a well-mixed
room). Consider a four-room house in which "B" denotes the bedroom, "L" denotes the living
room, "K" denotes the kitchen, and "P" denotes the porch (Figure 1). Here, the subscript "BL"
denotes the bedroom-to-living room air movement, "LB" reflects living room-to-bedroom air
movement, and HOB" denotes outdoors-to-living room air movement. Thus, the air flow rate from
the bedroom to the living room is wBL (volume of air per unit time, or L3/T), and the reverse air
flow rate is w^. The outdoors-to-bedroom air flow rate is wOB, which reflects the air passing
through the windows from outdoors or seeping through cracks in the walls. If the ambient air
2
-------
95-WP84B.03
has pollutant concentration x0, then the quantity of pollutant carried into the bedroom from
outdoors over time T will be the integral from t = 0 to / = T of the product of the concentration and
the outdoors-to-indoors flow rate, or the integral of xQwOB over time T (mass per unit time).
Assume that the initial concentration in the room is xB = 0 at time t = 0. Using the mass
balance approach, each integral in Equation 1 is the total quantity of pollutant either entering or
leaving the room's air from time t = 0 to time t = T. The sum of these quantities equals the amount
of pollutant assumed to be present in the room at time T, or xgvB:
T T T T T T
jgdt * jx0w0Bdt + j\wudf - jxBwBOdt - jxBwBLdt - jkjcgdt = *flvB (1)
Source Outdoor Living Room Outdoor Living Room Deposition Bedroom
(Cigarette) Infiltration Infiltration Exfiltration Exfiltration Sink Contents
The first three integrals on the left side of this equation, each with positive signs, denote the
quantity added to the room during period T. Of the three, the first integral is the quantity emitted
from a source within the room at emission rate g; the second is the quantity entering from
outdoors at concentration xQ and air flow rate wOB, and the third is the quantity entering from the
living room at concentration xL and air flow rate The fourth and fifth integrals, each with
negative signs, denote the quantity of pollutant lost from the room's air to outdoor air (integral of
xBwBO) and to the living room (integral of xBwBL). Finally, the sixth integral on the left-hand side
(integral of kjcB) represents the amount of pollutant lost to "sinks" in the room (for example, the
"plating out" of particles onto indoor surfaces). We assume that the deposition rate kB is
proportional to the concentration xB\ for pollutants such as CO that have no indoor sinks, kB = 0.
Setting the sum of the all integrals equal to the amount of pollutant present inside the room xBvB
(for bedroom volume vB) implies that mixing takes place very rapidly causing the concentration xB
inside the room to be uniform (approximately everywhere the same at any time 7).
Differentiating Equation 1 and rearranging terms, we obtain:
dx
Vb1T + x^BO + Wbl + **) = g + X°W°B + (2)
Because the total air entering the bedroom must equal the total air leaving the bedroom —
that is, there is no compression or leakage — we can substitute wBO + wBL = wOB + and then
divide both sides of the equation by wOB + + kB giving:
w
OB
+ w,
LB
+ ka
dt
W,
** =
WOB + WLB+ kB
+
OB
W,
WOB + VLB + ^B
+ X,
LB
WOB + WOL + ^B
(3)
3
-------
95-WP84B.03
Using XB(s) to represent the Laplace transform8 of the bedroom time series and XG(s) to represent
the Laplace transform of the source term g, the ratio of the Laplace transforms X^syX^s) is the
system transfer function for the source emission rate g relative to the bedroom concentration xa:
XJs) 1 I "Vm + wrj> * kg
Tr B(s) = = where . = a)
08 *G(*> vB , + p rk ** y * *
X£s) vL s +
L
In general, the system transfer function from any external room A to any adjoining room B
is written as follows:
XJs) w „ 1 [sum of air flows into room] ~ k„
Tab^ = tttt = T- where $b * (6)
XP) VB S + ¦ VB
Using the general form given by Equation (6), we can write the transfer function for the living
room "L" to the kitchen "K" simply by inspection of Figure 1. For example, the sum of all the air
flows into the kitchen is given by + wOK + wPK, yielding the following transfer function:
XJs) wrk. i vffJ. + w™ + w.„ + kr
% (7)
System transfer functions allow the analyst to predict the time series of concentrations in
all the rooms of the house from the time series of the source g = g(t) if its Laplace transform G(s)
is known. Application of these transform techniques to the house requires, of course, the values
of the parameters wBL, wOL, v£, etc., as well as the air exchange rate for each room. Even without
the values of these parameters, the inverse Laplace transforms of a time series can be used to
determine the general shapes of the concentration time series plots.
Consider a single cigarette smoked in the bedroom from time t = 0 to time t = 7 minutes
and a residence time (reciprocal of the air exchange rate) of an hour or more. Using the delta
function 6(f) as an idealized representation of a cigarette with total emissions of qagmg, the
Laplace transform of the cigarette will be XG(s) = 1 • q^. For a single cigarette, therefore, the
4
-------
95-WP84B.03
Laplace transform of the bedroom concentration is the product of the cigarette's Laplace transform
and the source-to-bedroom transfer function TaB(s) given by Equation 4:
w ¦ - u (»)
Referring to a table of Laplace transforms6, the concentration predicted in the bedroom as a
function of time from the inverse Laplace transform is given by the following exponential
function:
X(t) = lnLe**' far t > 0 (9)
v*
Similarly, the Laplace transform for the concentration predicted in the living room from a single
cigarette smoked in the bedroom is given by the product of Equation 5 and Equation 8:
XL(s) - Xt(s)TBL(s) = X^s)rjs)T,p) - -L^- -I- (I0)
Referring to a table of inverse Laplace transforms6, we find that the concentration time series
predicted for the living room from the cigarette in the bedroom is given by:
= —~ f°r ' - °' ^ (n)
vbvl(4>l-
This function begins at the origin, since xfi(0) = 0. Differentiating Equation 11 and setting the
result equal to zero shows that this function has a single mode xmax at time t = t„
'mac
r = _Z2£_E e I and t = (12)
vbvl<&l"
All the transfer functions for the compartments of a house are of the form in Equation 10; the
denominators contain the products (s + /4)(5 + Thus, all the solutions ~ the time series in all
rooms of the house ~ will consist of the sums of exponential functions when the source is a single,
relatively short "pulse" such as a cigarette. Representing the cigarette as a short pulse is an
approximation; if the residence time is short relative to the duration of the cigarette, then the
cigarette should be represented by a "rectangular" input function, as we have done elsewhere7.
5
-------
EXPERIMENTS IN A HOME
95-WP84B.03
To evaluate whether these equations adequately predict the time series of concentrations
for a cigarette smoked in an unoccupied home, over 40 experiments were conducted in a single-
story, two-bedroom house with outside dimensions of 28 ft by 24 ft, or 672 ft2 (62,5 m2) on
Partridge Ave. in Menlo Park, CA. The house was awaiting a new tenant, and we could vary the
locations of the monitors and the door and window positions in a variety of configurations.
Spatial Variation within a Bedroom
An important assumption of the mass balance model (Equation 1) is that, for any time t,
the room is sufficiently well-mixed to give nearly the same concentration at all locations. To
evaluate this assumption, experiments were conducted at three widely spaced locations in the
907,5 ft3 (25,7 m3) bedroom: (1) near the floor in the corner of the room, (2) on a short step
ladder 36" high in the center of the room; and (3) on a tall ladder 8.5" from the ceiling (Figure 2).
CO concentrations were measured using a Langan L15 CO Personal Exposure Measurer1 (Langan
Products, San Francisco, CA). Wires ran from each CO sensor to a DataBear8 data logger, and
precision electronic operational amplifiers multiplied the signals by 10.0 to increase the
sensitivity. CO concentrations in parts-per-ten-million (pptm) were logged at 30-second intervals.
Over a period of 16 hours, three Marlboro regular filter cigarettes were smoked in the
center of the bedroom at a 36" height with the doors closed and one window partly open but
covered with a shade. The first cigarette was smoked just after noon (12:46:30 pm) and lasted for
6 minutes and 30 seconds; the exponential decay of the CO concentrations from this cigarette was
quite similar for all three locations and lasted approximately 4.25 hours until just before 5:00 pm
(Figure 3). The second cigarette began approximately at 5:00 pm and lasted for 7 minutes and 16
seconds; it generated an exponential decay curve at each of the three locations, although the
concentration at the corner floor was lower than the concentration time series at the center of the
room or at the ceiling. Finally, a third Marlboro cigarette was smoked at 9:56 pm for 9-1/2
minutes, and its exponential decay was observed until 3:49 am. Each time series followed
Equation 9. The air exchange rate was determined by subtracting the background concentration
and taking the logarithms; the slope of the resulting straight Jine yielded approximately the same
ventilatory air exchange rate of 4>v = 1.2 air changes per hour (ach) at all locations.
The CO concentration time series, when averaged over each smoking episode, ranged from
3.88 pptm to 5.9 pptm, with an average for all episodes of 5.4 pptm (Table I). The center of the
room averaged about 1-1.2 pptm (21-23%) higher than the corner floor and about 0.7-1.3 pptm
(19-33%) higher than the ceiling. One would expect a higher average concentration in the center,
because the cigarette was smoked there within 12" of the CO monitor. It is unlikely that a person
would spend several hours either at the two extreme locations — the corner floor or the ceiling —
but would move about the room, so the person's average exposure would deviate by less than the
19-33% difference observed for the three locations. The average exposure of a person inside the
room probably would be closer to the overall mean of 5.4 pptm.
6
-------
95-WP84B.03
Table 1. Average CO Concentration (pptm) Measured in the Bedroom at Three Locations
after Smoking of Three Successive Marlboro Regular Filter Cigarettes
Experiment No.
Comer Floor
Center of Room
Top of Ladder
Mean
1
5.62
6.81
5.90
6.11
2
5.18
6,30
5.61
5.70
3
4.19
5.16
3.88
4.41
Mean:
5.00
6.09
5.13
5.41
Baughman et al? studied a chamber with 40 sampling points to determine how rapidly the
concentrations at different points converge, and Mage and Ott10, in reviewing their work, suggest
three time phases: an cc-period in which the source is emitting and the concentrations vary
spatially, a P-period in which the source is off but the room is not well-mixed, and a y-period in
which the coefficient of variation across all points is less than 0.10. There are too few locations in
the bedroom experiment to compute the coefficient of variation meaningfully, but Figure 3 shows
that the concentration time series pattern after the cigarette ends are similar at the three locations.
For a cigarette, the a-period is short relative to the other periods.
Our published cigarette smoking time series model7 derives the the following expression
for the average concentration over time T as a function of the average source strength g(T), the
instantaneous concentration x(T), the volume v, and the air exchage rate 4>:
isaLfJS.. . SI (13)
At the end of the exponential decay period, x(7) » 0 so that
Am"Se^S= 1.2 air changes per hour (ach), v = 25.7 m3, and the
average experiment duration of 4.25 hours, we can compute the total source emissions for the
cigarette experiment in the bedroom:
7
-------
95-WP84B.03
q = (p.541 ppm)[ 1.145 mg If 1.2 °ir C*a"ge 1(4.25 hrllil ^ ] = 81.2 mg
' m -ppm )\ hr ) air change)
The resulting total CO emission of 81.2 mg is not too different from the total CO emission of 88
mg obtained by combining mainstream and sidestream smoke from Marlboro cigarettes from
experiments in a chamber and an automobile and is similar to values reported in the literature.7
Notice that, at time / = 0, the exponential function in Equation 9 gives x(0) = q„Jvb= 81.2 mg/25.7
m3 = 3.16 mg/m3 CO. This result converts to (3.16 mg/m3)(l ppm-m3/1.147 mg)(10 pptm/ppm) =
27.6 pptm, which is very close to the peak concentration observed in Figure 3.
Concentrations in Two Rooms
In another experiment, the windows were closed, the door into the living room was
opened, and three Kentucky reference cigarettes No. 2R1 were smoked one after another in the
bedroom to obtain a strong source. RSP concentrations were measured (2-min averages) in the
living room and center of the bedroom using a Model 8510 piezobalance (TSI, Inc., St. Paul, MN).
The CO sensor on the tall ladder (location No. 3 in the bedroom) was moved into the living room
from the bedroom, but the other two bedroom sensor locations were unchanged (Figure 2). PAH
concentrations were measured at the corner floor location using a real-time PAH monitor
(EcoChem Technologies, Inc., West Hills, CA) that has been used in other ETS experiments."
Although the RSP concentrations for the two rooms initially diverge, they rapidly come
together after 45 minutes and remain very similar for the next four hours (Figure 4). Thus, with
the door between the two rooms open, the two rooms act almost as a single compartment. The
RSP concentrations caused by the three research cigarettes were extremely high, reaching a peak
of 5,500 ng/m3 in the bedroom. Indeed, the levels were so high that the three investigators found
it necessary to open a living room window at 4:00 pm, and the effect of opening this window in
changing the air exchange rate is especially evident in the CO time series plots (Figure 5). Unlike
CO, both RSP and PAH plate out on surfaces, and PAH therefore shows a more rapid decay than
CO in the bedroom.
Concentrations in Three Rooms
To examine the relationships among the time series in three rooms, a cigar was smoked in
the kitchen to serve as a strong source, and the resulting CO concentrations were measured in all
three rooms. Long coaxial cables were extended into each room with Langan LI5 CO sensors
attached, and CO concentrations on all three channels were logged using a DataBear* data logger.
As before, precision operational amplifiers multiplied the voltages by 10 to increase the
8
-------
95-WP84B.03
REFERENCES
1. P.L. Jenkins, T.J. Phillips, E.J. Mulberg, and S.P. Hui, "Activity Patterns of Californians:
Use of and Proximity to Indoor Pollutant Sources," Atmos. EnvironVol. 26A, No. 12,
pp. 2141-2148(1992).
2. T.E. McKone, "Household Exposure Models," Toxicology Letters, Vol. 49. pp. 321-339
(1989).
3. C.R. Wilkes, M.J. Small, J.B. Andelman, N.J. Giardino, and J. Marshall, "Inhalation
Exposure Model for Volatile Chemicals from Indoor Uses of Water," Atmospheric
Environment, Vol. 26A, No. 12, pp. 2227-2236 (1992).
4. J.W. Axley and D. Lorenzetti, "IAQ Modeling Using STELLA™: A Tutorial
Introduction," Building Technology Program, Massachusetts Institute of Technology,
(Fall 1991).
5. L. E. Sparks, B. A. Tichenor, J. B. White, J. Chang, and M.D. Jackson, "Verification and
Uses of the Environmental Protection Agency (EPA) Indoor Air Quality Model," Paper
No. 91-62.12 presented at the ZAth Annual Meeting of the Air and Waste Management
Association, British Columbia, June 16-21 (1991).
6. J.D'Azzo and C.H. Houpis, Feedback Control System Analysis and Synthesis, McGraw-
Hill (New York 1966).
7. W. Ott, L. Langan, and P. Switzer, "A Time Series Model for Cigarette Smoking Activity
Patterns: Model Validation for Carbon Monoxide and Respirable Particles in a Chamber
and an Automobile," J. Exposure Anal, and Environ. Epidemiology, Vol. 2, Suppl. 2, pp.
175-200(1992).
8. L. Langan, "Portability in Measuring Exposure to Carbon Monoxide," J Expos. Anal, and
Environ. Epidemiology, Supplement 1, pp. 223-239 (1992).
9. A. V. Baughman, A.J. Gadgil, and W.W. NazarofF, "Mixing of a Point Source Pollutant
by Natural Convection Flow within a Room," Indoor Air, Vol. 4, pp. 114-122 (1994).
10. D.T. Mage and W. R. Ott, "The Correction for Nonuniform Mixing in Indoor
Microenvironments," paper presented at the Symposium on Methods for Characterizing
Indoor Sources and Sinks, ASTM, Washington, DC (1994).
11. W.R. Ott, N.K. Wilson, N. Klepeis, and P. Switzer, "Real-time Monitoring of Polycyclic
Aromatic Hydrocarbons and Respirable Suspended Particles from Environmental
Tobacco Smoke in a Home," Proceedings of the International Symposium, "Measurement
of Toxic and Related Air Pollutants," Air and Waste Management Association, Omni
Durham Hotel, Durham, NC, May 3-6 (1994).
Preceding page blank
-------
Outdoors
95-WP84B.03
XqWob
i
4-
Xq.Wql
XB,Wbo
t
*
Xl^lo
Living Room v,
xB,wBL—~
\ <—XlWlb
Cigarette Source
Bedroom vB
XL-WLK
¦it
XK>WKL
Kitchen vK
xK'wko-
Xk^kp
It
Porch vp
xp, wpk
Xp, Wpo'
•XqWok
'X0,W0p
Figure 1. Schematic showing air flow paths and pollutant
concentrations in a 4-room house for which the
time series model was developed.
II
-------
IS9
PAH
Monitor
Piezo-
balance
Miniram
CO Monitor
VERTICAL HEIGHTS:
1. RSP, Miniram, PAH: 16"
CO: 5"
2. RSP, CO, Miniram: 36"
3. CO: 95" (8.5" from ceiling)
RSP: 90"
NO
Figure 2. Monitoring Locations in Bedroom of Partridge Avenue House.
00
w
o
-------
to
50
E 40
Q.
Q.
o
m
»-
«4-»
c
0
o
c
o
O
O
O
30
20
10
I
12:00 PM 3:20 PM 6:40 PM 10:00 PM 1:20 AM
1:40 PM 5:00 PM 8:20 PM 11:40 PM 3:00 AM
Time
11/11/93 to 11/12/93
Backgrounds Subtracted
Exp iv, v, vi
corner center ladder
Figure 3. CO Concentration measured in a 25.7 m3 bedroom over a 16-hour period
during which 3 Marlboro regular filter tip cigarettes were smoked.
Ul
I.
£
w
"o
u»
-------
95-WP84B.03
Concentration (ug/m3)
6,000
RSP RSP
Liv. Rm, Bedroom
5,000
4,000
3,000
Living room window opened at 4:00 PM
2,000
1,000
2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00 6:30
Time (PM)
Figure 4, RSP concentrations measured in the bedroom and living room with the door
open after three Kentucky No. 2R1 reference cigarettes were smoked.
Concentration (ppm CO, ug/m3 PAH)
12
10
8
6
4
2
°2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00 6:30
Time (PM)
Figure 5. CO and PAH concentrations measured in the bedroom and the living room with the
door open after three Kentucky No. 2R1 reference cigarettes were smoked.
Cigarettes started at 2:12:25 and
... — smoked consecutively until 2:20:30
;** :
l H
CO
j I Living room window
Bedroom
: u V.. opened at 4:00 PM
PAH
•/' \
Bedroom
1
Sr * \ •
CO
i vV 1
Liv. Rm.
I ?"• N .
1 v -.
14
-------
UI
£5 100
CO
£3 60
0
11:30 PM
Kitchen
Living Room
Bedroom
12:45 AM
2:00 AM
Time
3:15 AM
Figure 6. The CO time series in three rooms of the house after a cigar was smoked in the kitchen. The kitchen door was open 3"
and the bedroom window aid door were closed. The background concentration has been subtracted from each CO monitor.
-------
t
95-WP84B.03
120
E
100
phi = 0.000257 /sec = 0.0154/min
R2 = .998
Q.
a
c
o
N = 1098
n
c
o
o
c
O
o
O
O
0
2,000 4,000 6,000 8,000 10,000 12,000 14,000
Time (seconds)
Figure 7. The CO concentration measured in the kitchen after a cigar was smoked
along with the exponential fit to the data. The kitchen door was open 3"
70
£*60
D.
3 50
c
o
« 40
2
§ 30
o
820
o
O 10
\ f] \ \\ j Kitchen ACM = 0.0154
! // j V; Living Room1 ACM = 0.08^
! 1
151 i 1
! •'/ 1 1SN jScalinj* Factor = 9.^ j
1 II 1 i W i Peak Delav =25 minutes i
: i
! .. i
meas liv rm
1 1 1
j: i
\\ i I " I j i
\\ i i i 1 i
| |
Dred liv rm
l! I
i! i
i i !
1 ^Siv i 1 i
1 .... 1
i 1
i*
•
i*
4
¦
I
I
1 - -
i rvsi ! !
1 i i i
i i 1^, Vi. i
! !
! i
i j
I I
J L
25 50 75 100 125 150 175 200 225 250
Time (minutes)
Figure 8. Comparison of measured and predicted levels of CO in the living room after a cigar was
smoked in the kitchen. The door between the kitchen and the living room was open 3".
16
-------
TECHNICAL REPORT DATA
jfUmt mud Imuucttmt m ike n>mt brfort compti
I. REPORT NO.
EPA/600/A-95/07#
s.
4. TITLE ANO SUBTITLE
"Modeling Environmental Tobacco Smoke in the Home
Using Transfer Functions"
|B. REPORT DATE
June 1995
IS. PERPORMINO ORflANHATlON CODS
7. AUTMQRISI
a. perporminq organization report no.
Wayne R. Ott, Neil E. Klepeis, and Paul Switzer
• ¦ PERFORMINO ORGANIZATION name ANO ADDRESS
SIMS - Dept. of Statistics
Sequoia Hall 126
Stanford University
Stanford, CA 94305
16. PROGRAM t If MINT NO.
U.CflhTmeT/fiiAfcT k&.
CAG CR 814694
12. SPONSORING AGENCY NAM! ANO AODRESS
Human Exposure and Field Research Division (MD-56)
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
1 J. TYPE OP REPORT ANO PCRiOO COVtRID
Peer-Reviewed Arttcle/Proceedinfes
14. SPONSORING AGENCY COOt
ti. SUPPLEMENTARY NOTiS
It. ABSTRACT
This paper presents the theoretical and practical development of a multi-compartment indoor
aii quality model designed for predicting pollutant concentrations from environmental tobacco
smoke (ETS) in the home. The model is developed using transfer functions for each
compartment, thereby obtaining analytical solutions that can be expressed mathematically and
do not require a computer. The input parameters to the model are the cigarette source
emission rate, smoking activity patterns, room volumes, compartmental air exchange rates,
and intercompartmental flow rates. Field experiments are conducted in an unoccupied home
using a cigar and cigarettes as sources to evaluate the performance of the model, and real-
time measurements are made in the home of carbon monoxide (CO), respirable suspended
particles (RSP), and polycyclic aromatic hydrocarbons (PAH). The time series predicted
firom the equations by the model agree well with the concentration time series measured in
the rooms of the home. The transfer function approach can be applied to any home amply
by inspecting the floor plan and then writing the transfer functions by following simple rules.
The experimental data show that die door and window positions in each room exert
considerable influence on the pollutant concentrations observed in the home.
17. KEY WORDS ANO OOCUMSNT ANALYSIS
i. DESCRIPTORS
b.tBiNTIPIERS/OPEN ENDED TERMS
c. COSATi Field Group
Exposure Modeling
Msthftoatlcal Modeling
Indoor Air Quality Modeling
Environmental Tobacco Smoke (ETS)
Indoor Air Quality
Carbon Monoxide
1«. DISTRIBUTION STATEMENT
i
RELEASE TO PUBLIC
If SECURITY CLASS iTiiu Report,
UNCLASSIFIED
51. NC 0* »AGIS
17
to SECURITY class 1 Tim pun
UNCLASSIFIED
22. P° Zi
SPA Pm 2220-1 (*•«. 4.77) PRtviou* coition >• oiioliti
------- |