Building and Environment 35 (2000) 579±585
www.elsevier.com/locate/buildenv
The in¯uence of an architectural design alternative (transoms) on
indoor air environment in conventional kitchens in Taiwan
Che-Ming Chiang*, Chi-Ming Lai, Po-Cheng Chou, Yen-Yi Li
Department of Architecture, National Cheng-Kung University, Tainan, Taiwan, ROC
Received 6 April 1999; accepted 24 May 1999
Abstract
This study intends to investigate indoor air environment via the ¯ow ®elds, temperature ®elds and air contaminant (carbon
monoxide) distributions in the conventional residential kitchens, and look for e€ective methods to solve those problems through
natural ventilation techniques. Numerical simulations of the physical problem under consideration have been performed via a
®nite volume method for solving the governing equations and boundary conditions. It is obvious that location of accumulation
of air contaminants is highly relevant to the location of gas ®res, and using the residential range hood will be successful in
eliminating air contaminants. An architectural design alternative, utilizing transoms, is proposed to improve indoor air
environment in kitchens. 7 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Indoor air environment; Ventilation; CFD; Transom; Residential kitchen
1. Introduction
Indoor air quality in Taiwan is poor due to crowded
living space, highly airtight buildings, poor air circulation, and lack of ventilation. This problem has been
receiving more and more attention. Furthermore, a
large fraction of typical indoor pollutants (oil-aerosol,
water vapor, carbon monoxide, nitrogen oxides, and
VOCs) are released, owing to conventional cooking
behavior in Taiwan, e.g. decocting, frying, boiling,
scrambling, etc. Chiang et al. (1996) performed in-situ
®eld measurements of the impact of outdoor air and
living behavior patterns on indoor air quality of apartments in Taiwan [1]. It appears that the highest average concentration of carbon monoxide in residential
kitchens occurred during cooking-period when there
was a large variation shown in the indoor concentration (24-h average), ranging from 0.1±13.9 ppm.
There were 25% cases whose indoor hourly concen* Corresponding author.
E-mail address: [email protected] (C.-M. Chiang).
tration exceeded 9 ppm, ASHRAE criteria. According
to statistics issued by Taiwan's Department of Health,
lung cancers ®gure highly among women's diseases in
Taiwan. It is suspected that a strong relationship exists
between lung cancers and indoor air quality in kitchens, due to the existence of oil-aerosol and gaseous
contaminants.
Our research group has concentrated on constructing the ®eld-measurement system and developing the
P.O.E. (Post-Occupancy Evaluation) method on
indoor air quality in Taiwan over the past several
years [2±6]. The citations listed in these representative
studies provide the background information for this
problem, but they provide few informative strategies
for improvement.
The objective of the present study is to provide
insights into architectural design alternatives that will
improve indoor air environment in the conventional
kitchens in Taiwan considered for which little or no information is available. In so doing, numerical simulations via a ®nite volume method have been
performed for the steady three-dimensional turbulent
¯ow phenomena induced by a heat source (gas ®re).
0360-1323/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 3 6 0 - 1 3 2 3 ( 9 9 ) 0 0 0 2 9 - 3
580
C.-M. Chiang et al. / Building and Environment 35 (2000) 579±585
Table 1
Geometric data of the model kitchen (unit: cm)
Model kitchen
Cooking bench
Gas ®re
Range hood
Interior opening
Transom
Location of the transom, X
Fig. 1. Schematic diagram of the model kitchen investigated.
Complementary to the numerical situations, full-scale
experiments were also conducted for the physical con®guration under consideration. Of particular emphasis
in this study were the window-transom proposals
designed to attain an improved indoor air environment
in kitchens.
2. Method
2.1. Model kitchen
In Taiwan, many houses are built by the government, and then sold to citizens. Usually we call them
`public houses'. In the processes of planning, designing, constructing, and maintaining, economic moduli
and threshold limits of necessities for living are taken
into consideration. So they are usually referred to as
`conventional buildings', to a great extent. Ho et al.
(1993) surveyed almost all the public houses in Taiwan
and then got conclusive information about furniture
types/layouts and conventional length scales as they
appeared in those public houses [7]. In accordance
with those research results, main length scales and relevant parameters for a model kitchen are then decided.
270 210 240
270 53 80
70 43 21.2
89 57.6 42
90 120 (elevation: 90)
30 80 (elevation: 210)
40, 80, 120, 160
It is depicted schematically in Fig. 1. The geometric
data of this model house is listed in Table 1.
2.2. Physical problem
Fig. 1 illustrates schematically the physical con®guration of the air environment undergoing ventilation
processes through transom opening (natural ventilation) or drawn by range hood (mechanical ventilation) in the conventional residential kitchens in
Taiwan. Initially air is at its domestic temperature Ti
in Taiwan. Suddenly the heat ¯ux generated by the gas
®re raises the temperature of air adjacent to this
region. The buoyancy-driven air¯ow resulting from the
temperature di€erence between the gas ®re and ambient air is assumed to be three-dimensional and turbulent. When we operate the range hood, forced air ¯ow
drawn by the exhaust fans in the range hood a€ects
natural convection ¯ow as described before. In such a
case, a `mixing convection' phenomenon occurs. Furthermore, the thermophysical properties of the air are
temperature independent, except for the density, for
which the Boussinesq approximation is valid. The
other dimensional parameters speci®ed in this study
are shown in Table 2.
Table 2
Dimensional parameters speci®ed in numerical calculation
Initial temperature Ti
Wall of the kitchen
Heat ¯ux generated by the gas ®re
Carbon monoxide produced by combustion
Outdoor carbon monoxide concentration
Suction ¯ow rate of the range hood
268C (statistically average temperature in summer in Taiwan)
Adiabatic
5.9 105 (W/m2)
1.1 10ÿ3(kg/m2 s)
3 ppm
530.385 m3/h
Table 3
Con®gurations for numerical simulation (NO means not open/operate)
Features
Range hood
Interior door
(1)
(2)
(3)
(4)
NO
NO
Operated
NO
Closed, but the lower louvers are still opened
Dominant ¯ow pattern
Pollutant accumulation
E€ect of range hood
E€ect of architectural alternative±transom
Exterior door
Interior opening
Transom
Opened
NO
NO
NO
NO
NO
NO
Opened
C.-M. Chiang et al. / Building and Environment 35 (2000) 579±585
581
Fig. 2. Velocity vector diagram (left), temperature distribution (middle) and CO concentration pro®le (right) (feature 1).
2.3. Numerical method
Numerical simulations of the physical problem
under consideration have been performed via a ®nite
volume method for solving the governing equations
and boundary conditions mentioned above. This study
applies the SIMPLE (Semi-Implicit Method for Press-
ure Linked Equations) algorithm [8] to solve those
equations. The `two equation model' of turbulence, the
k-epsilon model [9], was adopted. To bridge the steep
dependent variable gradients close to the solid surface,
the `general wall function' is employed. The iteration
calculation was continued until a prescribed relative
convergence of 10ÿ4 was satis®ed for all ®eld variables
582
C.-M. Chiang et al. / Building and Environment 35 (2000) 579±585
Fig. 3. Velocity vector diagram (feature 2).
Fig. 5. CO distribution (feature 2).
of this problem. A 38 28 32 grid system was
employed for the present calculations.
steady-state, three-dimensional turbulent ¯ow in nine
conventional kitchens, with the following parameter
con®gurations listed in Table 3. The numerical results
will therefore be presented with a primary focus on the
in¯uence of those con®gurations on the air¯ow patterns, temperature ®elds and carbon monoxide distributions in those kitchens.
2.4. Full-scale experiment
In order to reveal the indoor air environment in the
kitchen, experiments were also performed in a fullscale model kitchen. (Built in the National ChengKung University) that mimics the physical con®guration depicted in Fig. 1. The characteristics of the air
velocity ®eld are measured by a three-dimensional
ultrasonic anemometer. The accuracy of the ultrasonic
anemometer is 22% within full scale (within the range
of the main air velocity direction after zero-point
adjustment). The temperature ®eld is measured by Ttype thermocouples and recorded in the data logger.
The uncertainty of temperature measurement is less
than 2%.
3. Numerical results
Numerical simulations have been undertaken for the
Fig. 4. Temperature ®eld (feature 2).
3.1. Dominant ¯ow pattern
The numerical results will therefore be presented
with a primary focus on the in¯uence of the gas ®re's
locations (left, middle, and right-hand side of the
bench) on turbulent ¯ow patterns and temperature/carbon dioxide distributions in the conventional kitchens,
as illustrated in Fig. 2. When the gas ®re is located on
one side of the bench, air contaminants with high temperatures will accumulate at the corner of the upper
space; when the gas ®re is located on the middle, all
the distributions are symmetric and the accumulations
still exist above the gas ®re. It is obvious that the location of the accumulation of air contaminants is
highly relevant to the location of the gas ®re. Accord-
Fig. 6. Velocity vector diagram (feature 3).
C.-M. Chiang et al. / Building and Environment 35 (2000) 579±585
Fig. 7. Temperature ®eld (feature 3).
ing to the ¯ow patterns we discovered, we decided
intuitively to `punch' a hole (transom) which can expel
the convective thermal-plume to the outdoors directly.
This is the context of those cases where the removal
eciency of the transom is examined.
583
Fig. 9. Velocity vector diagram (feature 4).
3.3. E€ect of range hood
Flow structures developed within the conventional
kitchen under uniform heat ¯ux (5.9 105 w/m2) are
presented by means of a velocity vector diagram, as illustrated in Fig. 3. The following plots display numerical result in the cross-section plane through the
location of a person cooking, unless explicitly stated
otherwise. The convective thermal-plume from the heat
source generated by the gas ®re causes an upwind air
movement. The warm air forms temperature and carbon monoxide strati®cation regions under the ceiling,
as illustrated in Figs. 4 and 5. Temperature and carbon
monoxide concentration around the person's breathing
zone are about 608C and 124 ppm respectively, a
dangerous situation for a person in such an environment [10].
When we operate the range hood, the suction of air
would intensify the upwind air movement. So, we can
observe that overall air¯ow structure is very similar to
the case when the range hood is turned o€, as illustrated in Fig. 6. Air with high temperature and contaminants can be exhausted directly to the outside
through the range hood. The predicted results around
the breathing zone are 27.68C and 4.5 ppm respectively, as displayed in Figs. 7 and 8. It is obvious that
using range hoods cannot only make carbon monoxide
concentration decrease (from 124±4.5 ppm) to an
acceptably healthy level (lower than ASHRAE criteria,
9 ppm) [11] but also make the temperature go down
(from 60±27.68C) to an acceptably comfortable level.
However, further e€orts on the natural ventilation
system have to be made, due to energy-saving aspects
(electrical consumption of the range hood) and noise
made by range hood operating. An architectural
alternative, transoms, are therefore proposed in this
study.
Fig. 8. CO distribution (feature 3).
Fig. 10. Temperature ®eld (feature 4).
3.2. Pollutant accumulation
584
C.-M. Chiang et al. / Building and Environment 35 (2000) 579±585
Fig. 12. Variations of temperature and CO concentration at di€erent
transom location, X.
Fig. 11. CO distribution (feature 4).
3.4. E€ect of architectural alternatives
Figs. 9±11 exemplify the ¯ow structures, temperature ®elds and carbon monoxide distributions in the
model kitchens when location of the transom, X, is
40 cm. Fig. 12 shows that variation of temperature
and carbon monoxide concentration in the model
kitchens occurs when the transom is installed at a
di€erent location. It appears that the transom's performance of removing high temperature air with contaminants is excellent, especially when the location is
properly arranged. We can realize that to attain an
improvement of indoor environment, not only depends
on mechanical facilities, but proper architectural design Ð another economical way.
4. Comparison with full-scale experiments
Finally, to further validate the accuracy of the numerical simulations undertaken, the predicted ¯ow patterns are compared with results of the ¯ow velocity
measurements, as illustrated in Fig. 13. Fig. 14 shows
the computed temperature ®eld with the measurements
for opened interior opening, constant air inward velocity from exterior door (0.3 m/s), constant inward air
temperature (26.48C and 14.88C) and the operated
range hood. A generally good qualitative agreement
can be readily observed.
5. Discussion
In the present article, the problem of improvement
of indoor air environment in the conventional residential kitchens in Taiwan has been studied numerically
by means of ®nite volume method. Results from the
numerical simulations undertaken indicate that:
1. The location of the accumulation of air contaminants is highly relevant to the location of the gas
®re in conventional kitchen in Taiwan.
2. Using the range hood will make elimination of air
contaminants (CO) e€ective.
3. Characteristics of ¯ow structure do not alter too
much whether the range hood is operated or not.
4. When the location of the transom is close to the
region of accumulation of air contaminants, its performance of removing air contaminants is excellent.
Fig. 13. Comparison of the predicted ¯ow pattern (left) with full-scale experiments (right).
C.-M. Chiang et al. / Building and Environment 35 (2000) 579±585
585
pecially grateful to ARCHILIFE Research Foundation
for their computational facilities.
References
Fig. 14. Variations of temperature at the breathing zone with the distance to the interior opening.
To verify the present simulations, full-scale experiments mimicking the physical con®guration considered have been conducted and generally good
agreement between the prediction and the experiments
was observed for the ¯ow patterns and temperature
®elds in the model kitchens.
Acknowledgements
Support from the National Science Council of ROC
through grant No. NSC 88-2211-E-006-047 in this
study is gratefully acknowledged. Also, we are es-
[1] Chiang CM, Chou PC, et al. A study of the impacts of outdoor
air and living behavior patterns on indoor air quality Ð case
studies of apartments in Taiwan. INDOOR AIR '96
1996;3:735±40.
[2] Chiang CM, Wang WA, Hu SC. A study of indoor air quality
in a naturally ventilated bedroom. Healthy Building/IAQ '97
1997;3:87±90.
[3] Chiang CM, Chou PC, Lai CM, et al. Numerical simulation of
natural ventilation of a bedroom in warm climate.
ROOMVENT '98 1998;2:563±7.
[4] Chiang CM, Lai CM, Chou PC, Li YY. Predicted improvement
of indoor air environment in the conventional residential kitchens in Taiwan. ICHES '98, p. 626±32 1998.
[5] Wang WA, Chou PC, Chiang CM. Investigation on the air¯ow
distribution in a cross-ventilated residential bedroom. Healthy
Building/IAQ '97 1997;3:87±90.
[6] Chung KC, Chiang CM, Wang WA. Predicting contaminant
particle distributions to evaluate the environment of lavatories
with ¯oor exhaust ventilation. Building and Environment
1997;32(2):149±59.
[7] Ho UF, et al. Development of the geometric guideline for civil
houses in Taiwan. Department of the Interior, Taiwan: Building
Research Institute, 1993 (in Chinese).
[8] Patankar SV. Numerical heat transfer and ¯uid ¯ow.
Washington: Hemisphere, 1980.
[9] Launder BE, Spalding DB. Computer methods in applied mechanics and engineering. North-Holland Publishers Co, 1974.
[10] ASHRAE Standard 55a, Thermal Environmental Conditions
for Human Occupancy. 1995.
[11] ASHRAE Standard 62-1989, Ventilation for Acceptable Air
Quality. 1989.