Performance rating of glass windows and glass windows with films in aspect of thermal comfort and heat transmission

Abstract

This article is about a study on glass window and glass window with film of different types in aspect of thermal comfort and heat transmission. Different types of glass window, clear glass, tinted glass, reflective glass, double pane glass, and low-e glass were investigated. Films with different spectral optical properties were then adhered to the glass windows of different types and studied. The analysis was done based on the outside design weather condition which selected from 12 years of Bangkok meteorological data. Predicted percentage of dissatisfied (PPD) was selected as the thermal comfort index. The relative heat gain (RHG) based on local weather condition was selected as the heat transmission index. The PPD can be subdivided into the PPD due to surface temperature effect and the PPD due to solar radiation effect. The analysis indicated that, for most of the glass windows considered except the reflective glasses, the values of PPD due to solar radiation effect were much larger than the values of PPD due to surface temperature effect. And the most discomfort condition occurred when using a clear glass as window. Adhered films to the glass windows caused the PPD due to surface temperature effect increase and cause the PPD due to solar radiation effect decrease. It was also found that the PPD values due to solar radiation effect for glass windows and glass windows with films were varied linearly with the total transmittance of glass windows and glass windows with films. The PPD values due to surface temperature effect were varied with the total absorptance of glass windows and glass windows with films in an almost linear fashion. The heat transmission index, RHG, based on chosen design weather condition can be subdivided into the RHG due to conduction effect and RHG due to solar radiation effect. The analysis indicated that the values of RHG due to solar radiation effect were larger than the values of RHG due to conduction effect for all glass windows and glass windows with films considered in this study. Adhered film to the glass windows resulted in lowering the relative heat gain due to solar radiation in the amount corresponding to the film properties. But the film had very few effect on the relative heat gain due to conduction. The relative heat gain values were varied linearly with the total transmittances of the glass windows and glass windows with films. The relative heat gain values were also varied inversely with the absorptances of glass windows and glass windows with films in a linear fashion.

Full text

Performance rating of glass windows and glass windows with
films in aspect of thermal comfort and heat transmission
Somsak Chaiyapinunt
a,
*, Bunyarit Phueakphongsuriya
a
,
Khemmachart Mongkornsaksit
a
, Nopparat Khomporn
b
a
Mechanical Engineering Department, Faculty of Engineering, Chulalongkorn University, Bangkok 10330, Thailand
b
Mechanical Engineering Department, Faculty of Engineering, Siam University, Bangkok 10160, Thailand
Received 18 May 2004; received in revised form 28 July 2004; accepted 4 October 2004
Abstract
This article is about a study on glass window and glass window with film of different types in aspect of thermal comfort and heat
transmission. Different types of glass window, clear glass, tinted glass, reflective glass, double pane glass, and low-eglass were investigated.
Films with different spectral optical properties were then adhered to the glass windows of different types and studied. The analysis was done
based on the outside design weather condition which selected from 12 years of Bangkok meteorological data. Predicted percentage of
dissatisfied (PPD) was selected as the thermal comfort index. The relative heat gain (RHG) based on local weather condition was selected as
the heat transmission index. The PPD can be subdivided into the PPD due to surface temperature effect and the PPD due to solar radiation
effect. The analysis indicated that, for most of the glass windows considered except the reflective glasses, the values of PPD due to solar
radiation effect were much larger than the values of PPD due to surface temperature effect. And the most discomfort condition occurred when
using a clear glass as window. Adhered films to the glass windows caused the PPD due to surface temperature effect increase and cause the
PPD due to solar radiation effect decrease. It was also found that the PPD values due to solar radiation effect for glass windows and glass
windows with films were varied linearly with the total transmittance of glass windows and glass windows with films. The PPD values due to
surface temperature effect were varied with the total absorptance of glass windows and glass windows with films in an almost linear fashion.
The heat transmission index, RHG, based on chosen design weather condition can be subdivided into the RHG due to conduction effect and
RHG due to solar radiation effect. The analysis indicated that the values of RHG due to solar radiation effect were larger than the values of
RHG due to conduction effect for all glass windows and glass windows with films considered in this study. Adhered film to the glass windows
resulted in lowering the relative heat gain due to solar radiation in the amount corresponding to the film properties. But the film had very few
effect on the relative heat gain due to conduction. The relative heat gain values were varied linearly with the total transmittances of the glass
windows and glass windows with films. The relative heat gain values were also varied inversely with the absorptances of glass windows and
glass windows with films in a linear fashion.
#2004 Elsevier B.V. All rights reserved.
Keywords: Glass windows; Films; Thermal comfort; Heat transmission; Performance rating
1. Introduction
Large office and commercial buildings in Thailand
usually have large amount of glass windows installed as the
building envelopes. The glass windows are installed to serve
as physical and visual connection to outsiders, as well as
to make the appearance of buildings look more aesthetic.
And since Thailand is located in the tropical zone near the
equator. The weather is hot and humid for most of the year.
Therefore, besides the advantage of the glass windows as
described above, the glass windows installed in buildings in
Thailand also act as a means to admit solar radiation into
buildings and convert it into building heat gain and then
building cooling load, respectively. Such buildings which
are air conditioned will usually consume more energy from
the air conditioning system to take care of the cooling load
due to large amount of solar radiation passing through glass
www.elsevier.com/locate/enbuild
Energy and Buildings 37 (2005) 725–738
* Corresponding author. Tel.: +66 22186610; fax: +66 22522889.
E-mail address: fmescy@eng.chula.ac.th (S. Chaiyapinunt).
0378-7788/$ – see front matter #2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.enbuild.2004.10.008

windows. At the same time, Thailand has also issued the
energy conservation promotion act B.E. 2535 (1992) [1] and
ministry regulation on designated building B.E. 2538 (1995)
[2] to control the energy usage in buildings. According to the
regulation, the amount of heat gain through building
envelopes (Overall Thermal Transfer Value, OTTV and
Roof Thermal Transfer Value, RTTV) are limited to certain
values. The regulation also enforced on the existing
buildings and new buildings. Large buildings in Thailand
that constructed before the regulation been issued (defined
as existing buildings) usually have large amount of glass
windows installed as building envelopes. And those
buildings shall have the values of heat gain passing through
envelopes exceeding the regulation values. Therefore those
buildings have to be corrected by changing the envelope
thermal properties to reduce the heat gain through the
envelopes to the regulation value. The easiest way to change
the building envelope properties especially on the glass
windows is to adhere films on the glass windows. Data of
film properties available for customer in Thailand are
usually given in the form of overall values (not in function of
wavelength) and the values given are usually referenced to a
clear glass. But glasses used for glass windows in existing
buildings can have various types, such as, clear glass, tinted
glass, and reflective glass, etc. Therefore, when one wants to
change the glass window properties to reduce heat
transmission by adhering the film to the glass window
which is not a clear glass, one cannot directly use values of
the given film properties to analyze. Spectral optical
properties of the individual glass and film are needed in
order to find the spectral properties of glass with film. Glass
windows and glass windows with films affected the building
not only on thermal transmission but they also affected on
thermal comfort and visual comfort. Therefore, the under-
standing of the thermal performance of glass windows and
glass windows with films in aspect of heat transmission and
thermal comfort shall be the essential things for design
architects, design engineers, building owners, and officers
who responsible for enforcing and issuing energy policy.
This article describes the study for thermal performance
rating of the glass windows and glass windows with films
under local design condition.
2. Thermal comfort index and heat transmission index
Thermal comfort is defined as the condition of mind that
expressed satisfaction with the thermal environment (ISO
7730 [3] and ASHRAE Standard 55 [4]). Thermal comfort
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738726
Nomenclature
a
k
absorptance of the outer surface of the person
(standard value = 0.6)
A
T
total absorptance
eemittance of glass
E
t
global solar radiation, W/m
2
f
cl
clothing area factor
f
p
projected area factor
F
p–i
angle factor between the person and surface i
(Pn
iFpi¼1)
h
c
convective heat transfer coefficient, k cal/
(h m
2
8C)
I
dirv
direct solar radiation striking on the vertical
glass surface, W/m
2
I
diffv
diffuse solar radiation striking on the vertical
glass surface, W/m
2
Mmetabolic rate per unit body, k cal/(h m
2
)
N
i
inward flowing fraction of absorbed radiation
P
a
vapor partial pressure, mmHg
PMV predicted mean vote
PPD predicted percentage of dissatisfied
qsolar radiation intensity passing through glass
window that the person in the enclosure
exposed, W/m
2
q
A
instantaneous rate of heat admission through
glass window and/or glass window with film,
W/m
2
Rh
k
resistance due to conductance, includes radia-
tive affects at layer, m
2
K/W
Rk
k
resistance due to solid conductance at k layer,
m
2
K/W
SC shading coefficient
SHGC solar heat gain coefficient
SHGF solar heat gain factor, W/m
2
t
i
inside temperature, 8C
t
j
temperature at center of glass at jlayer, 8C
t
o
outside temperature, 8C
t
sj
surface temperature of the enclosure wall num-
ber j,8C
t
sk
surface temperature of glass at ksurface, 8C
T
a
Air temperature, 8C
T
cl
Clothing surface temperature, 8C
T
herm
hemisphere spectral glass transmittance
T
mrt
mean radiant temperature, 8C
T
smrt
mean radiant temperature due to surface
temperature and solar radiation, 8C
T
tmrt
mean radiant temperature due to surface
temperature, 8C
T(u) glass transmittance which dependent on inci-
dent angle
Uoverall coefficient of heat transfer, W/(m
2
8C)
Greek letters
aabsorptance of glass window and glass window
with film
e
p
emittance of the outer surface of the person
(standard value = 0.97)
hmechanical efficiency
sStefan–Boltzmann constant, W/m
2
K
4
t
T
total transmittance

can be predicted in several ways. One way to describe the
state of thermal comfort is to use predicted mean vote
(PMV) index and predicted percentage of dissatisfied (PPD)
index. PMV index is defined as the index predicting the
mean response of a large group of people to thermal
environment. The index is classified as seven levels between
+3 (hot) to 3 (cold) which level 0 is neutral. PPD index is
defined as the percentage of people who will not be satisfied
with the thermal environment they occupied. These indices
take into account six parameters which affected the thermal
state: activity, clothing, air temperature, mean radiant
temperature, air velocity, and humidity. The values of
PMV range of 0.5 which corresponding to the PPD value
of 10% are usually accounted as acceptable condition.
Fanger [5] related PMV value to the parameters that affect
the thermal comfort and also related the PPD value to the
PMV value by the following equations:
PMV ¼ð0:352 e0:042Mþ0:032Þ½Mð1hÞ
0:35ð43 0:061Mð1hÞPaÞ
0:42ðMð1hÞ50Þ0:0023Mð44 PaÞ
0:0014Mð34 TaÞ3:4108fclððTcl þ273Þ4
ðTmrt þ273Þ4Þfcl hcðTcl TaÞ (1)
PPD ¼100 95 eð0:03353PMV4þ0:2179PMV2Þ(2)
where M= metabolic rate per unit body, k cal/(h m
2
),
P
a
= vapor partial pressure, mmHg, f
cl
= clothing area fac-
tor, T
mrt
= mean radiant temperature, 8C, T
a
= air tempera-
ture, 8C, T
cl
= clothing surface temperature, 8C,
h
c
= convective heat transfer coefficient, k cal/(h m
2
8C),
h= mechanical efficiency.
One of the important variables needed to solve for PMV
in Eq. (1) is the mean radiant temperature. Mean radiant
temperature (MRT) of an enclosure is defined as that
uniform temperature of an imaginary black enclosure which
would result in the same heat loss by radiation from the
person as the actual enclosure. When one is doing the study
of the thermal environment of a person who sit near a large
glass window, besides the glass surface temperature that
affect the room mean radiant temperature, one has to include
the effect of solar radiation beam that striking on the person
too. Because sunlight can easily penetrate through glass
window striking directly on the occupants inside which
causes a much higher mean radiant temperature. Unfortu-
nately the expression of PMV in Eq. (1) cannot be able to use
for directly calculating PMV that include the effect of solar
radiation. Therefore, in order to take into account of the
effect of solar radiation striking on the person in the
enclosure, two mean radiant temperatures shall be used in
this study. The first is the mean radiant temperature of the
enclosure that does not account for solar radiation (some-
times called unirradiated mean radiant temperature). This
mean radiant temperature is mainly dominated by glass
surface temperature. The second mean radiant temperature
is the total mean radiant temperature which accounted for
the effect from surface temperature and solar radiation. Such
mean radiant temperatures can be written as
Ttmrt ¼½ðts1þ273Þ4Fp1þðts2þ273Þ4Fp2þ 
þðtsn þ273Þ4Fpn0:25 273 C (3)
Tsmrt ¼ðTtmrt þ273Þ4þfpak
q
eps

0:25
273 C (4)
where T
tmrt
= mean radiant temperature due to surface tem-
perature, 8C, T
smrt
= mean radiant temperature due to sur-
face temperature and solar radiation, 8C, t
sj
= surface
temperature of the enclosure wall number j,8C, F
p–i
= angle
factor between the person and surface i(Pn
iFpi¼1),
f
p
= projected area factor, a
k
= absorptance of the outer sur-
face of the person (standard value = 0.6), e
p
= emittance of
the outer surface of the person (standard value = 0.97),
s= Stefan–Boltzmann constant, W/m
2
K
4
,q= solar radia-
tion intensity passing through glass window that the person
in the enclosure exposed, W/m
2
The solar radiation passing through glass window qcan
be determined by the following relation as:
q¼IdirvTðuÞþIdiffv Therm (5)
where I
dirv
= direct solar radiation striking on the vertical
glass surface, W/m
2
,I
diffv
= diffuse solar radiation striking
on the vertical glass surface, W/m
2
,T(u) = glass transmit-
tance which dependent on incident angle, T
herm
= hemi-
sphere glass transmittance.
As previously stated, Eq. (1) can be used to calculate only
for the PMV that does not have the solar radiation effect, the
PMV which is also accounted for the solar radiation effect
shall be calculated by using the relation suggested by Lyons
[6] and Sullivan [7] as:
dPMV
dq¼@PMV
@Tmrt
@Tmrt
@ðakfpqÞ
@ðakfpqÞ
@q(6)
PMV ¼PMVno solar þdPMV
dqq(7)
Then PPD values that accounted for the solar radiation effect
and surface temperature effect can be calculated from
Eq. (2) by using PMV values obtained from Eq. (7). Then
the value of PPD due to solar radiation effect alone can be
obtained by subtracting the PPD value due to surface
temperature effect from total PPD value according to the
relationship shown in Eq. (8).
PPDðtotalÞ¼PPDðsurface temperatureÞ
þPPDðsolar radiationÞ(8)
The glass window surface temperature can be calculated by
using the method suggested by Finlayson et al. [12]. The
calculation is done based on balancing the heat flux of each
layer of glass and environment. The calculation is performed
with the glass system of Nlayers and 2Nsurfaces and also
involved with the iteration process to yield the converging
solution. The expression of the surface temperature can be
written as:
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738 727

For the outermost surface
ts1¼ðt1=Rk1to=Rh1Þ
ð1=Rk1þ1=Rh1Þ(9)
and
tsN ¼ðtN=RkNti=RhNÞ
ð1=RkNþ1=RhNÞ(10)
For the remaining surfaces of the jlayers of the glazing
system where surface 2jfaces inward and surface 2j1 faces
outwards, the surface temperature can be written as:
tsð2j1Þ¼ðtj=Rkjt2j=Rhjþ1Þ
ð1=Rkjþ1=Rhjþ1Þ(11)
and
ts2j¼ðtjþ1=Rkjþ1t2j1=Rhjþ1Þ
ð1=Rkjþ1þ1=Rhjþ1Þ(12)
where t
sk
= surface temperature of glass at ksurface, 8C,
t
j
= temperature at center of glass at jlayer, 8C, t
o
= outside
temperature, 8C, t
i
= inside temperature, 8C, Rk
k
= resis-
tance due to solid conductance at klayer, m
2
K/W, Rh
k
= re-
sistance due to conductance, includes radiative affects at
layer, m
2
K/W.
Heat transmission through glass windows and glass
windows with films can be broken into components as solar
heat gain and conduction heat gain. And the solar heat gain
can be further broken into components as solar radiation
transmitted through glass and inward flow of absorbed solar
radiation. They can be expressed as:
qA¼EttþNiðaEtÞþUðtotiÞ(13)
where q
A
= instantaneous rate of heat admission through
glass window and/or glass window with film, W/m
2
,
E
t
= global solar radiation, W/m
2
,N
i
= inward flowing frac-
tion of absorbed radiation, a= absorptance of glass window
and glass window with film, U= overall coefficient of heat
transfer, W/(m
2
8C), t
o
= outside temperature, 8C, t
i
= inside
temperature, 8C.
Sometimes this instantaneous heat gain can be written by
substituting t+N
i
awith solar heat gain coefficient or
SHGC as:
qA¼EtSHGC þUðtotiÞ(14)
The instantaneous heat gain in Eqs. (13) and (14) can be
rewritten in the simplify form as:
qA¼SC SHGF þUðtotiÞ(15)
where SC = shading coefficient, SHGF = solar heat gain
factor, W/m
2
.
Since the purpose in this study is to find the performance
rating for windows based on the same reference condition,
therefore the index chosen shall be the instantaneous heat
gain in Eq. (15) and called as relative heat gain (RHG). The
heat transmission index then can be expressed as:
RHG ¼SC SHGF þUðtotiÞ(16)
3. Validation of some mathematical models
The experimental results chosen to validate the simulated
heat gain through glass windows are the measured values
performed by Klems et al. [13] and the measured values
performed by Morya [16]. The first set of experimental
results are the measured values performed on the MoWiTT
(Mobile Window Thermal Test) facility which the unit
located on the campus of the University of Nevada, Reno.
The MoWiTT is a mobile calorimetric facility designed to
measure the net heat flow through a fenestration as a
function of time under realistic outdoor conditions. The
details of the MoWiTT is described by Klems [14] and
Klems et al. [15]. The measured values chosen for this study
are the measured heat gain through two specific glass
windows. The first glass window chosen is the clear
uncoated double glazing with air-filled gap. The second
glass window chosen is the clear coated double glass with
argon-filled gap. The outer pane glass is coated with low-e
(e<1) coating on the inside surface. The glass windows
selected utilized frameless sealed-insulating glazing unit
with two panes of 3 mm glass separated by 12.7 mm gas
filled gap. All units are 0.91 m wide 1.22 m high.
Without exact information on the type of low-ecoating
used on the second set of glass window (e <1), the low-e
glass window with the following optical properties;
transmittance as 0.429, emittance as 0.042, reflectance
front as 0.338 and reflectance back as 0.418, is chosen as the
outer pane glass for validated purpose. The simulated results
on heat gain for both glass windows with the same inside and
outside condition on the tested day are shown with the
experimental results in Fig. 1. There was a correction for the
time lags between the driving solar flux and the calorimeter
response to yield the correct instantaneous net heat flow.
Twenty minutes lag between the calorimeter heat flow
measurement and the incident intensity was assumed as
suggested by Klems et al. [13]. The simulated results agree
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738728
Fig. 1. Comparison between the heat gain through selective double glass
windows from measurement on MoWiTT and from the simulation.

very well with the experimental results especially on clear
glass. Small discrepancy on the maximum value of the heat
gain on the coated glass is observed. This discrepancy is
suspected to come from the use of the coated emittance as
0.042 in the simulation which could be different from the
real coated used in the tested glass. The second set of
measurement values obtained from the work done by Morya
[16] at AIT (Asean Institute of Technology, Pathumthani,
Thailand). The measurement was performed at the facility
on the campus by measuring the air inside and outside
temperature, glass inside and outside surface temperature,
solar radiation on direct normal, diffuse, and global
component that incident on glass surface and transmitted
solar radiation. The air temperature inside and outside of the
tested room, and glass surface temperature of the tested
room were measured by type K thermocouple. The
transmitted and reflected solar radiation through the room
was measured by solarimeter. The measurement in this case
was not performed with calorimetric room concept as
MoWiTT. The inside condition in the tested room was not
controlled. The data were instantaneous collected and then
calculated to obtain the heat gain by using mathematical
model. The experimental results from Morya [16] were
expected to have less accuracy when compared to the results
from MoWiTT. The measured results chosen for this study
which done on different days are the heat gain per unit area
through a 6 mm clear glass window and through a 6 mm
tinted glass window. The simulated results under same
operating condition and using same mathematical model for
inside and outside surface film coefficient as ones used in the
results obtained from measured data are shown in Fig. 2 with
the measured results. The trend of the heat gain from the
simulated results and the measured results are in the same
pattern. Some discrepancies are found. These discrepancies
are expected to be the effect of error in the measurement as
mentioned by Morya [16].
Most of the mathematical models used for calculating
thermal comfort parameters came from the empirical
formulae adopted from many experimental works by Fanger
[5]. One of the parameters that has major effect on the
thermal comfort index in this study would be the mean
radiant temperature especially the one due to surface
temperature and solar radiation (T
smrt
). The experimental
results on the thermal comfort index due to high solar
radiation passing through the glass windows are quite
limited. In this study, the results measured by Sullivan [7]
are chosen to perform the validation task. Sullivan [7] has
conducted some measurement in a room with solar radiation
passing through the glass window. The measurement values
are operative temperature, air temperature, and solar
radiation. The solar radiation was measured by Epply
pyranometer. The operative temperature was measured by
the comfort meter. The air temperature was measured by the
climate analyzer. Then the mean radiant temperature due to
surface temperature and solar radiation in a room was then
determined using the measured operative and measured air
temperature with the air velocity of 0.15 m/s by using the
relation given in ISO 7730 [3]. The mean radiant
temperature due to surface temperature for this case was
assumed to be equal to air temperature. The data are given in
Table 1. Then the mean radiant temperature due to surface
temperature and solar radiation (T
smrt
) is then calculated by
using the relation in Eq. (4) setting the air velocity as 0.15 m/
s, and the projected area factor as 0.3. The simulated results
and the experimental results on the mean radiant tempera-
ture are shown in Table 1. The difference in values are within
10% which could be considered acceptable for most general
experiment. The difference could be come from the
uncertainty and some error from measurement. With these
results of validation, the mathematical models for determin-
ing glass window performance can be use in this study with
confidence.
4. Design condition
Many thermal properties of glass windows and glass
windows with films, such as overall coefficient of heat
transfer (U), solar heat gain factor (SHGF), solar heat gain
coefficient (SHGC), and shading coefficient (SC), are
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738 729
Fig. 2. Comparison between the heat gain through selective glass windows
from measurement at AIT facility and from the simulation.
Table 1
Measured room temperature, operative temperature and solar radiation and
mean radiant temperature from measured data and from the simulation
Room
temperature
(8C)
Operative
temperature
(8C)
Solar
radiation
(W/m
2
)
MRT from
operative
temperature (8C)
MRT from
simulation
(8C)
21.5 30.0 562 38.5 38.05
21.5 31.3 588 41.1 38.76
23.0 31.8 603 40.6 40.43
21.0 30.2 577 39.4 38.04
21.0 31.1 596 41.2 38.57
21.0 31.5 598 42.0 38.61

dependent on the outside and inside condition of the
considered enclosure. For this study, the design outside
condition is chosen based on 12 years (1988–1999) of
Bangkok weather data collected by the meteorological
department. The selection is done based on considering
the most influencing parameters on the thermal perfor-
mance and thermal comfort of the glass windows and
glass windows with films. Such parameter is the solar
radiation. The 0.4% annual cumulative frequency of
occurrence for global radiation as suggested by ASHRAE
[8] is selected. The design outside condition is then
chosen as:
direct normal solar radiation on glass windows and glass
windows with films = 658 W/m
2
;
diffuse solar radiation on glass windows and glass
windows with films = 111 W/m
2
;
outside air dry bulb temperature = 35 8C;
outside wind velocity = 3.8 m/s.
The room that used for the study shall be 4 m 4m
and 3 m height with one external wall and three internal
partitions. There shall be glass window or glass window with
film installed the whole area of the external wall and facing
west. The person in this study shall sit turning sideway to the
glass window with a distance of 1 m. The typical working
condition for office in Bangkok are chosen as design con-
dition. The inside design condition in this study is then
chosen as:
inside air dry bulb temperature = 25 8C;
inside air velocity = 0.15 m/s;
relative humidity = 50%;
clothing insulation = 0.5 clo;
metabolic rate for activity = 1.2 met (1 met = 58 W/m
2
).
5. Glass and film
To accurately analyze the thermal performance of the
glass windows and glass windows with films, one needs to
know the spectral properties of the individual glass and glass
with film. Therefore the data of spectral properties which
varied with the wavelength are required. The solar radiation
shall have spectral range from about 0.38 to 3.5 mm. The
range of wavelength of the solar radiation spectrum can be
divided into the visible range (0.38–0.76 mm) and the
infrared range (0.76–3.5 mm). Different types and thickness
of glass are chosen in this study. They are clear glass, tinted
glass, reflective glass, double pane glass, and low-eglass
(low emittance coated glass, e= 0.04). Clear glass chosen
has different thickness ranged from 2 to 19 mm. Tinted
glasses chosen are bronze glass, gray glass, and green glass.
Reflective glasses are clear and tinted glasses coated with
reflective coating which gave the total reflectance values
around 60% and the total transmittance values less than
10%. They are referred as reflective clear glass, reflective
bronze glass, reflective gray glass, and reflective green glass.
Double pane glass chosen in this study are composed of
outside glass and inside glass which are separated by 6 mm
air gap. The outside glasses of the double pane glasses
chosen in this study are clear glass, bronze glass, gray glass,
green glass, reflective clear glass, reflective bronze glass,
reflective gray glass, reflective green glass, low-eclear glass,
low-ebronze glass, low-egray glass, and low-egreen glass
while the inside glass is a clear glass. The data of spectral
properties of the glasses and films chosen are obtained from
the library of program OPTIC 5 [9].
The spectral properties of glass and film, the transmit-
tance, absorptance, and reflectance, are the main properties
that affect on the thermal performance of glass windows and
glass windows with films. But since the summation of the
three properties (transmittance, absorptance, and reflec-
tance) is equal to one. One property of glass will affect the
other two (i.e. glass with high transmittance will have low
absorptance and reflectance). With the limitation of space
for this article, the transmittance is chosen to be the main
spectral property to represent the characteristics of glass and
film. The detail of the other two spectral properties of glasses
and films are listed in Ref. [10].Fig. 3 shows the normal
incidence spectral transmittances of clear glass with
different thickness. Fig. 4 shows the normal incidence
spectral transmittances of 6 mm clear and tinted glasses.
Fig. 5 shows the normal incidence spectral properties of
6mm reflective clear glass. Fig. 6 shows the normal
incidence spectral transmittances and reflectances of 6 mm
reflective glasses which illustrated the effect of different
reflective glass types on the spectral transmittances and
reflectances of the reflective glasses. Fig. 7 shows the normal
incidence spectral transmittances of some double pane
glasses. Fig. 8 shows the normal incidence spectral
properties of double pane low-eglasses. Some nomencla-
tures used in Figs. 4–8are as the following: CL = clear,
BR = bronze, GY = gray, GN = green, RCL = reflective
clear, RBR = reflective bronze, RGY = reflective gray,
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738730
Fig. 3. The comparison of the normal incidence spectral transmittance of
clear glass for different thicknesses.

RGN = reflective green, LCL = low-eclear, LBR = low-e
bronze, LGY = low-egray, and LGN = low-egreen.
From Fig. 3, the effect of glass thickness on the
transmittance is clearly seen in the infrared range, the thicker
the glass the less transmittance value it shall have, while the
glass thickness has less effect on the transmittance in the
visible range. From Fig. 4, the effect of color in the tinted
glass on the transmittance are shown. It is found that tinted
glasses have lower spectral transmittance values for the
whole spectrum ranges considered compared to the clear
glass of the same thickness. And the green glass shall have
largest value of spectral transmittance in the visible range
and lowest value in the infrared range among the tinted
glasses considered. That means the green glass will allow
more light transmittance into the room than the other two
types of tinted glasses. At the same time the green glass will
allow lower value of energy in the infrared range transmitted
into the room compared to the other two types of tinted
glasses considered. The typical characteristics of the
reflective glass which its normal incidence spectral pro-
perties are high reflectivity in the front surface, high
absorptivity, and low transmittivity are shown in Fig. 5. The
reflective glass shall have some values of transmittances in
the visible range and almost negligible values of transmit-
tance in the infrared range. Fig. 6 shows the spectral
transmittances and reflectances for different types of
reflective glasses. All reflective glasses chosen have lower
spectral transmittances in the visible and infrared range with
the reflective green glass shall have largest values of
transmittances in the visible range. The values of spectral
reflectances in the front surface of the three types of glasses
are about the same. The values of spectral reflectances in the
back surface of the three types of glasses are different as
shown in Fig. 6. Double pane glasses considered in this study
are composed of the 6 mm outer pane glass and 6 mm inner
clear glass with 6 mm air gap. The spectral transmittances of
the double pane glasses will have the similar pattern as the
single glasses of the same kind except lower in magnitudes
as can be seen in Fig. 7. The low-eglass is the double pane
glass which are consisted of clear glass and tinted glasses
applied with low-ecoating on the inner surface of the outer
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738 731
Fig. 4. The comparison of the normal incidence spectral transmittance of
clear glass and tinted glasses of 6 mm thickness.
Fig. 5. The normal incidence of different spectral properties of typical
reflective clear glass of 6 mm thickness.
Fig. 6. The normal incidence spectral transmittance and reflectance in the
front and back surface of the tinted reflective glasses. Tr refers to transmit-
tance, Rf refers to reflectance of the front surface of the glasses, Rb refers to
reflectance of the back surface.
Fig. 7. The normal incidence spectral transmittance of different types of
double pane glasses.

pane glass separated from the inner pane glass with 6 mm air
gap. The typical characteristic of low-eglass chosen is that it
has a lower value of spectral transmittance in the infrared
range while allow more light transmitted in the visible range
as can be seen in Fig. 8. One can see from Fig. 8 that the
reflectance of the front surface of the glasses considered are
about the same while the reflectance of the back surface of
the glasses considered are different. The transmittance of
double glass is lower than a single glass of the same kind.
Four types of film (types A–D), which have different
visual appearances and different spectral characteristics in
their properties, are chosen for this study. In each type of the
chosen film, films which have different amounts of
transmittance are also included in the study. Therefore, 14
different films are used for the study. Fig. 9 shows the
comparison of the spectral transmittance for different kinds
of specific chosen film which has approximately the same
amount of transmittance in the visible range and also shows
the comparison between the spectral transmittance for the
same type of film with different amount of transmittance in
the visible range. These films shall be adhered to the inside
glass surface which facing toward the room. The spectral
optical properties of glass window that is not a single glass
and properties of glass windows with films are determined
by method suggested by Rubin et al. [11].
6. Analysis
The analysis is done by firstly obtaining the spectral
properties of individual single glass of different types and
thickness and the spectral properties of double pane glass of
different types. Then the spectral properties of glass windows
with films of different types are obtained by combining the
spectral properties of individual glass and film by using
OPTIC 5 [9] which the calculating method is based on the
model suggested by Rubin et al. [11]. Then the total optical
properties (in solar range (wave length 0.32–2.5 mm)) of glass
windows and glass windows with films can be obtained by
doing the integration of the product of spectral properties and
the spectral irradiance over the assigned wavelength range
and weight with the result from the integration of the spectral
irradiance itself over the same assigned wavelength range
(0.32–2.5 mm is chosen to be limit of integration). With the
known optical property data of the glass windows and glass
windows with films, their thermal properties, such as overall
heat transfer coefficient, shading coefficient, solar heat gain
coefficient, and glass surface temperature that based on
Thailand design weather condition, are calculated by using
method suggested by Finlayson et al. [12]. Then, with the
calculated glass surface temperature, the mean radiant
temperature due to surface temperature and total mean
radiant temperature can be calculated. The variation of PMV
on the mean radiant temperature is calculated by determining
the increment on the PMV when increasing the mean radiant
temperature by 1 8C as suggested by Fanger [5]. The other two
derivative terms in Eq. (6) are then determined as the finite
difference between each related term in the solar exposed
condition and no solar exposed condition. The thermal
comfort index (PPD) and the heat transmission index (RHG)
shall then be calculated. The results of the analysis for some
typical glass windows are shown in Figs. 10 and 11. Some
nomenclatures used in Figs. 10 and 11 are as the following: (1)
3 mm clear, (2) 6 mm clear, (3) 9 mm clear, (4) 12 mm clear,
(5) bronze, (6) gray, (7) green, (8) reflective clear, (9)
reflective bronze, (10) reflective gray, (11) reflective green,
(12) clear_clear, (13) bronze_clear, (14) gray_clear, (15)
green_clear, (16) reflective clear_clear, (17) reflective
bronze_clear, (18) reflective gray_clear, (19) reflective
green_clear, (20) low-eclear, (21) low-ebronze, (22) low-
egray, and (23) low-egreen.
From the results of thermal comfort index values (PPD)
shown in Fig. 10, one can conclude that most of the glass
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738732
Fig. 8. The normal incidence spectral properties of low-eglasses which
composed of clear and tinted glass with low-ecoating on the outer pane and
clear glass as the inner pane with 6 mm air gap. Tr refers to transmittance, Rf
refers to reflectance of the front surface of the glasses, Rb refers to
reflectance of the back surface.
Fig. 9. The comparison between the normal incidence spectral transmit-
tance of different types of film. The first letter refers to type of the film
chosen (types A–D) and the second number refers to the amount of
transmittance of the film in the visible range.

windows except reflective glass shall have the values of PPD
due to solar radiation greater than the values of PPD due to
surface temperature. The reason is that most of glass
windows considered in this study shall have values of their
transmittances greater than their absorptances. Therefore the
effect of solar transmission through glass window is greater
than the effect of solar absorption in the glass (which caused
the glass surface temperature to increase). Only the
reflective glasses chosen, which have greater values in
reflectance and absorptance than transmittance, shall have
the values of PPD due to surface temperature greater than the
values of PPD due to solar radiation. Clear glass shall have
values of total PPD higher than the other types glass
windows. The single reflective glasses chosen shall have the
lowest values of total PPD among the single glasses of
different types. As for double pane glasses, the double pane
glass which has clear glass as the inside and outside pane
shall have the values of total PPD highest among the double
pane glasses considered. The double pane glass which has
reflective glass as the outer pane shall have the values of total
PPD lowest. And the PPD due to solar radiation is greater
than the PPD due to surface temperature for most of glasses
considered except the ones with the reflective glass as the
outer pane. The double pane glass shall have lower values in
total PPD than of the single glass of the same type.
For heat transmission study, the result of the analysis for
each glass window and glass window with film shall be
composed of the effect of conduction and effect of solar
radiation. The solar heat gain factor under Bangkok design
weather data is chosen to be 430 W/m
2
. Most of the glass
windows considered, the effect of solar radiation heat gain
on the heat transmission index is far greater than the effect of
conduction heat gain as shown in Fig. 11.
The results of heat transmission analysis for glass
windows show that clear glass has the greatest values of the
relative heat gain among the glasses studied. The reflective
glass windows have the smallest value in relative heat gain.
Double pane glass has smaller value in relative heat gain
compared to the single glass of the same type. That is,
double pane glass composed of clear glasses have the
greatest value of relative heat gain while the double pane
glass composed of reflective glass as the outer pane have the
smallest value in relative heat gain.
As for glass windows with films in this study, the films
chosen shall only be applied to the clear, tinted, and double
pane clear glass of 6 mm thickness. The spectral properties
of the glass windows with films will be corresponding to the
combination of the spectral properties of the individual glass
(Figs. 4 and 7) and film (Fig. 9) itself. The normal incidence
of spectral transmittances of some glass windows with films
are shown in Figs. 12–16. Some nomenclatures used in
Figs. 12–19 are as the followings: the first letter refers to
type of the film chosen; A = type A film, B = type B film,
C = type C film, D = type D film, the second number in the
legend means the amount of transmittance of the specified
film in the visible range; 2 = film with transmittance in the
visible range of 0.2, 5 = film with transmittance in the visible
range of 0.5, and the last two letters refer to the type of glass
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738 733
Fig. 10. The values of the thermal comfort (PPD) of different types of glass
windows.
Fig. 11. The comparison between the heat transmission index values for
different types of glass windows.
Fig. 12. The normal incidence spectral transmittance of clear glass adhered
with four different types of film.

that the films adhered; CL = clear glass, BR = bronze glass,
GY = gray glass, GN = green glass, C2 = double pane clear
glass. All glasses adhered with type A film shall have the
smallest value of the overall spectral transmittance compare
to glasses adhered with other types of film considered. The
spectral transmittance of double clear glass adhered with
films are in the same pattern with single clear glass with
films but have the smaller values. Among the tinted glasses
with films considered the green glass have the highest value
of the transmittance in the visible range. Then, the thermal
comfort index and heat transmission index for glass
windows with films can be obtained and some of the results
are shown in Figs. 17–19.Fig. 17 shows the effect of films on
the thermal comfort condition by comparing the thermal
comfort index of a bare glass with glass window with films.
Every film will cause the values of PPD due to surface
temperature to increase and cause the values of PPD due to
solar radiation to decrease. The amount of PPD values
changed by the film is primary related to the spectral
properties of the film applied.
The results of heat transmission analysis for glass
windows with films are shown in Figs. 18 and 19. It is found
that the films that applied to the glass windows can reduce
relative heat gain values when compared to the bare glass
window of the same kind. The amount of reduction in heat
gain is dependent on the characteristics of film itself. The
lower spectral transmittance the film has, the lower value of
heat gain through the window it shall be.
Finally all the results obtained are grouped together and
analyzed to find some relationship between the thermal
comfort index and properties of glass window and glass
window with film and between the heat transmission index
and properties of glass window and glass window with film.
And to be able to use the results from this study for practical
application in glass window thermal performance rating, the
total optical properties of glass windows are chosen as
the representing parameters instead of the spectral proper-
ties. Fig. 20 shows the relationship between the thermal
comfort index in the solar part (PPD(solar)) and the total
transmittance (sometimes called solar transmittances) of
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738734
Fig. 13. The normal incidence spectral transmittance of bronze glass
adhered with four different types of film.
Fig. 14. The normal incidence spectral transmittance of gray glass adhered
with four different types of film.
Fig. 15. The normal incidence spectral transmittance of green glass adhered
with four different types of film.
Fig. 16. The normal incidence spectral transmittance of double pane clear
glass adhered with four different types of film.

glass windows and glass windows with films in a linear
fashion. The relationship can be divided into two groups.
The first group is the single pane glasses, single pane glasses
with films, double pane clear glass, and double pane clear
glass with films. The second group is the double pane glasses
with the outer pane glasses are tinted glasses, reflective
glasses, and low-eglasses. Therefore the following relation-
ships between the PPD(solar) and glasses transmittance can
be expressed as the following:
PPDðsolarÞ¼96:807tT0:6517 (17)
PPDðsolarÞ¼75:167tT2:8995 (18)
where t
T
is the total transmittance (solar transmittance) of
the glass windows and glass windows with films.
Eq. (17) is applied to the first group of glasses and
Eq. (18) is applied to second group of glasses as mentioned
above. One can see that the glass windows and glass
windows with films shall give more discomfort condition
due to solar beam striking the occupant when the total
transmittance is increasing.
For another thermal comfort index which is the PPD due
to surface temperature (PPD(surface temperature)), its
relationship with the total absorptance of the glass windows
and glass windows with films can be shown in Fig. 21. In this
case the relationship can be divided into three groups. The
first group is the single pane glasses and single pane glasses
with films. The second group is the double pane glasses. The
third group is the double pane glasses with films. Therefore
the relationship between the PPD(surface temperature) and
glass absorptance can be expressed as the following:
PPDðsurface temperatureÞ¼5:1988 e1:5063AT(19)
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738 735
Fig. 17. The comparison between the values of the thermal comfort (PPD) of different types of glass windows adhered with films and bare glass windows.
Fig. 18. The comparison between the heat transmission index values for different types of glass windows adhered with films and bare glass windows.

PPDðsurface temperatureÞ¼3:258ATþ5:4373 (20)
PPDðsurface temperatureÞ¼45:284AT10:119 (21)
where A
T
is the total absorptance (solar absorptance) of the
glass windows and glass windows with films.
Eq. (19) is applied to the first group of glasses, Eq. (20) is
applied to the second group of glasses, and Eq. (21) is
applied to the third group of glasses. From Fig. 21, one can
see that the larger values of the total absorptance of glass
windows chosen the more discomfort condition from the
effect of glass windows surface temperature will be except
the bare double pane glass windows. One more interesting
point is that when applied film to the double pane glass
windows the effect of discomfort due to surface temperature
will be increased with the rate higher than the single pane
glass windows.
With the relationship of thermal comfort index and
properties of glass windows and glass windows with films as
shown in Figs. 20 and 21, one can find the total thermal
comfort index (state of discomfort) for any types of glass
windows and glass windows with films under Thailand
design weather condition for the purpose of the performance
rating to choose the proper glass windows and glass
windows with films in aspect of thermal comfort.
For the heat transmission study, the relationship between
the heat transmission index (relative heat gain) and the
properties of the glass windows and glass windows with
films can be obtained and displayed in Figs. 22 and 23.
Fig. 22 shows the relationship between the relative heat gain
values and the total transmittance of the glass windows and
glass windows with films. Fig. 23 shows the relationship
between the relative heat gain and the total absorptance of
the glass windows and glass windows with films. One can
conclude that the heat transmission through the glass
windows and glass windows with films is primarily
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738736
Fig. 19. The comparison between the heat transmission index values for different types of glass windows and double pane glass windows adhered with films and
bare glass windows.
Fig. 21. The relationship between the PPD due to surface temperature and
the total absorptance of glass windows and glass windows with films.
Fig. 20. The relationship between the PPD due to solar radiation and the
total transmittance of glass windows and glass windows with films.

dependent on the total transmittance of the windows. The
glass windows and glass windows with films with larger
values of the total transmittance shall have the larger values
of heat transmission. And in Fig. 23, one can see that for
most of the single and double glasses with and without films
which do not have special characteristics in high reflectances
shall have the relationship of heat transmission index values
and the total absorptance in a linear fashion with negative
slope. That means glass windows with high transmittance
and low absorptance values shall have heat transmission
more than glass windows with low transmittance and high
absorptance values. While the relationship of the heat
transmission index of the reflective glass windows with the
total transmittance and absorptance values are not quite
unique. Since the properties that have the major effect on
thermal performance of the reflective glass windows is the
reflectance.
7. Conclusion
The study shows that the thermal performance of glass
windows and glass windows with films are dependent on the
spectral properties of the glass windows and glass windows
with films. The thermal comfort index chosen (PPD) will
consist of PPD due to solar radiation effect and PPD due to
surface temperature effect. Most of the glass windows and
glass windows with films considered except reflective glasses
have larger values of PPD due to solar radiation than PPD due
to surface temperature. It is found that those glass windows
which have rather high values of transmittance (which shall
allow high solar radiation beam striking person in an
enclosure) shall cause more discomfort to a person in an
enclosure. On the otherhand, the reflective glasses which have
rather high absorptance values and low transmittance values
will receive more heat absorbed in glassmaterial than the heat
transmitted through theglass windows. The heat absorption in
the glass material will cause the glass surface temperature to
rise up significantly. Therefore PPD due to surface
temperature of the reflective glasses are larger than the
PPD due to solar radiation. The reflective glasses chosen in
this study are primarily concentrated on the reflection of the
energy out of the glass window not really concerned about
light transmission. Therefore if one chose the reflective glass
with considerable amount of light transmission, the thermal
performance in aspect of comfort and heat gain may be poorer
than the low-eglasses. Films when adhered to the glass
windows shall cause the transmittance of the glass windows
with films decreased and cause the absorptance of the glass
windows with films increased. Therefore glass windows with
films shall have the values of PPD due to solar radiation
decreased and the values of PPD due to surface temperature
increased when compared to plain glass windows. Though
adhered film to glass windows will cause the total PPD
decreased but it causes the PPD due to surface temperature
increased. Care has to be taken when one wants to use films
adhered to glass windows to reduce the heat gain. Also films
will cause certain reduction in light transmission through
glass windows with films. The spectral optical properties of
different types of films should be carefully studied to be able
to obtain the proper film to suit the required application. The
heat transmission index, RHG, is the index chosen for
compared the thermal performance in aspect of heat
transmission for different types of glass windows and glass
windows with films under the same local design condition. It
is found that the reflective glasses of single pane and double
panes have the smallest values of RHG compared to all glass
windows studied. Adhered films to glass windows cause the
reduction in heat transmission dependent on the properties of
the films chosen. It is also found that the total transmittance
and total absorptance of the glass windows and glass windows
with films shall be regarded as the key properties for
establishing the thermal performance in terms of thermal
comfort and heat transmission. Anyone who wants to choose
the glass windows and glass windows with films for using as
building envelopes should carefully consider their perfor-
mances in terms of heat transmission, thermal comfort, light
transmission, and appearance. And the results of this study
can be used as a guideline to do the proper choosing.
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738 737
Fig. 22. The relationship between the relative heat gain values and the total
transmittance of glass windows and glass windows with films.
Fig. 23. The relationship between the relative heat gain and the total
absorptance of glass windows and glass windows with films.

Acknowledgements
The authors are grateful for financial support from the
National Metal and Materials Technology Center, National
Science and Technology Development Agency.
References
[1] Energy Conservation Promotion Act B.E. 2535, Department of Energy
Development and Promotion (DEDP), Ministry of Science, Technol-
ogy and Environment (MOSTE), Bangkok 10400.
[2] Royal Decree on Designated Building B.E. 2538, Ministerial Regula-
tion issued under the Energy Conversation Promotion Act B.E. 2535,
Department of Energy Development and Promotion (DEDP), Ministry
of Science, Technology and Environment (MOSTE), Bangkok 10400.
[3] ISO 7730, Moderate thermal environments—determination of the
PMV and PPD indices and specification of the conditions for thermal
comfort, 1995.
[4] ASHRAE, Thermal Environmental Conditions for Human Occupancy,
ANSI/ASHRAE Standard 55-1992, 1992.
[5] P.O. Fanger, Thermal Comfort Analysis and Applications in Environ-
mental Engineering, McGraw-Hill, New York, 1970.
[6] P.R. Lyons, Window performance for human thermal comfort, ASH-
RAE Transactions (2000) 594–602.
[7] R. Sullivan, Thermal Comfort Issues in the LRI Study, Internal
Memorandum/Report, Windows and Day lighting Group, Lawrence
Berkeley National Laboratory, Berkeley, CA, USA, 1986a, 1986b.
[8] American Society of Heating, Refrigerating and Air Conditioning
Engineers, ASHRAE Handbook of Fundamental 2001, Chapter 27,
pp. 27.1–27.71.
[9] OPTIC 5, http://windows.lbl.gov/materials/optic5/ [cited 9 December
2003].
[10] S. Chaiyapinunt, K. Mongkornsaksit, B. Pheukphongsuriya, N. Kom-
porn, Analysis of indices for glass windows and glass windows with
film for buildings under Thailand weather condition in aspect of
thermal comfort and thermal energy transmission, Final Report,
Mechanical Engineering Department, Chulalongkorn University,
National Metal and Materials Technology Center, under contract
MT-B-44-CER-09-093-G, November 2002.
[11] M. Rubin, K. von Rottkay, R. Powles, Window optics, Solar Energy 62
(1988) 149–161.
[12] E. Finlayson, D. Arastech, C. Huizenga, M. Rubin, S. Reilly, Window
4.0 Documentation of Calculation Procedures, Lawrence Berkeley
Laboratory Report 33943, August 1993.
[13] J.H. Klems, Yazdanian, G.O. Kelly, Measured performance of selec-
tive glazings, in: Thermal Performance of Exterior Envelopes of
Building IV Conference Proceedings, Clearwater Beach, FL, Decem-
ber 4–8, 1995.
[14] J.H. Klems, Measurement of fenestration net energy performance:
considerations leading to development of the mobile window thermal
test (MoWiTT) facility, Journal of Solar Energy Engineering 110
(1988) 208–216.
[15] J.H. Klems, S. Selkowitz, S. Horowitz, A mobile facility for measuring
net energy performance of windows and skylights, in: Proceedings of
the CIB W67, Third International Symposium on Energy Conservation
in Built Environment, vol. III, Dublin, Ireland, AuForas Forbartha,
1982.
[16] K. Morya, A study on heat gain through windows in hot climate,
Thesis, Master of Science, Asean Institute of Technology, School of
Environment, Resources and Development, Pathumthani, Thailand,
August 1998.
S. Chaiyapinunt et al. / Energy and Buildings 37 (2005) 725–738738