Thermal Comfort Assessment in Two Tropical Regions and Radiant
Cooling as a Passive Cooling Option
Surapong Chirarattananon∗, Nyi Nyi Htan, Rizwan Ahmed Menon, and
Energy Field of Study, Asian Institute of Technology, P.O. Box 4, Klong Luang,
Field assessments of thermal comfort were conducted in Mehran University, in
Pakistan, and in the Asian Institute of Technology (AIT), north of Bangkok. In the
first case, students in classrooms ventilated by ceiling fans were the subjects in a
summer survey. It was found that the neutral effective temperature was close to
values predicted by adaptive models. The study in AIT included subjects in air-
conditioned classrooms and offices and subjects in student cafeteria ventilated by
ceiling fans. For those in air-conditioned classrooms, neutral effective temperature
was found to be close to 25oC, and for those in fan ventilated space the neutral
effective temperature was close to 28°C. The latter value was predicted closely by
adaptive models. The results do not differ significantly between subjects of different
nationalities, or genders in the AIT assessment. But the age of a subject show some
influence. Radiant cooling using passive means to generate cooling water for supply
to cooling panels on ceiling and wall of a room was studied experimentally at AIT,
and simulation results show that such system could be applied in both locations to
achieve natural comfort.
Keywords thermal comfort, air-conditioning, natural ventilation, neutral
temperature, thermal sensation.
Human beings would not only strive to have their dwellings safe and secure, they
would also like to make them comfortable. Thermal comfort (TC) contributes
substantially in making a dwelling comfortable. The subject of TC has been
associated with air-conditioning in the early history of the study of both subjects.
Results of TC studies using human subjects in air-conditioned environment
culminates in the development of thermal comfort standards such as ASHRAE/ANSI
Standards 55-2004 and ISO7730. The standards embodies a hypothesis that human
beings, regardless of race, age, sex, feel comfortable, or thermally neutral, in a well-
defined but narrow range of thermal environment. Air-conditioning (AC) is energy
intensive, but it is justified by an assertion that people in thermally neutral
environment work more productively. Lomonaco and Miller (1997) reported that
productivity gain of up to 17% had been achieved for clerical employees with
provision of appropriate air-conditioning. In warm and humid climate, AC for
cooling accounts for 60% of electricity consumption in commercial buildings,
Vangtook and Chirarattananon (2005). Air-conditioning has reached or nearly
reached saturation in large cities in hot humid cities such as Bangkok, Jakarta, Kuala
Lumpur, Manila, and Singapore. It is also increasingly used in residential households
Author for correspondence, email: email@example.com
where it typically contributes over 70% of electricity consumption. In motor cars, AC
is even more common. It now seems that AC is synonymous with TC in Thai society.
Recently, the Thai government mounted a campaign for energy conservation in the
midst of rising oil price. It advised its citizen to set the temperature in the households
and in the workplaces to 25°C. For daytime, air temperature almost always stays
above 25°C. Air-conditioning must be used to be able to set the air temperature in
dwelling to 25°C. The subject of TC is not well understood, and AC is often
associated with wealth and prestige. A household with air-conditioner installed in the
living room may not normally turn it on, but will turn it on when guests visit.
Field studies of TC with people residing in naturally ventilated (NV) dwellings
found that people would be thermally comfortable or neutral in a much wider range
of environment then those specified by ASHRAE Standard 55-2004 or ISO 7730,
where AC is assumed used to control the thermal environment, Nicol and Humphreys
(2002). Field studies of TC in NV dwelling in hot and humid climate have been
conducted in Hawaii, Kwok (1998), in Bangkok, Busch (1992), Yamtraipat etal
(2005), Khedari etal (2001), Jitkhajornwanich (2002), in Jogjakarta, Feriadi and Wong
(2004), in Singapore, Wong and Khoo (2003) and in Zambia, Malama etal (1998), all
produce results that show that thermally comfortable condition varies over a wide
range and is related to climate of a location. Similar to the results of Busch, the study
conducted in AIT illustrates that a person may feel TC in NV environment at
temperature higher than when he or she is in an AC environment. The two
environments clearly differ. To create the condition that enables a person to feel TC
requires much less cooling and thus less energy than that which is needed by an
environment defined by ASHRAE Standards 55-2004 or ISO7730.
This paper reports the results of field studies of TC at the Asian institute of
Technology (AIT), north of Bangkok, Thailand, and at Mehran University near
Hyderabad, Pakistan. It also reviews studies on application of radiant cooling
conducted at AIT, as a viable mode of passive cooling to be used to achieve natural
2. FIELD SURVEYS
Field surveys for assessment of thermal comfort were conducted in Mehran
University and in AIT. In both cases, each subject was requested to give personal
information that included clothing, food taken earlier, and activities undertaken before
the interview. In AIT, some female subjects refused to give information on clothing,
so observation was made instead. In Mehran University where female students were
fewer, most refused to be interviewed, so the subjects comprise male students only. In
the thermal assessment, seven-point sensation sale of ASHRAE and three-point Mc
Intyre scale were used. Values of physical variables of the environment surrounding
respondents were measured by a set of instruments classified as class II field
instruments according to Brager and de Dear (1998). This set of field instruments
measured air temperature, relative humidity, air speed, and radiant temperature. The
instruments were put on a tray supported by a stand to 1 meter height from the floor.
This corresponded to the height of the shoulder of a subject at his sitting position.
Each subject was given a sheet of questionnaire that the subject marked or filled
his/her short answers.
2.1 AIT Survey (NV and AC spaces)
The Asian Institute of Technology was established to provide international
postgraduate education where the medium of instruction has been English. It does not
offer undergraduate degree program. The subjects in the survey were mostly Asian,
but there are exchange students from European and African countries, totally of 20
nationalities. The ages of the subjects ranges from early twenties to over fifty. The
campus is located fourty kilometers north of Bangkok. The latitude of the location is
14oN. The climate is hot and humid. Table 1 shows some statistical values of
temperature and relative humidity of air. Mean wind speed is low, at 1.5 m/s.
Table 1 Monthly mean values of temperature and relative humidity of air.
Quantity Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Mean max. 31.9 32.8 33.9 34.9 34.2 33.1 32.6 32.4 32.0 31.8 31.5 31.4
Mean 25.6 27.2 28.6 29.6 29.3 28.7 28.1 27.9 27.6 27.5 26.7 25.5
Mean min. 20.6 23.1 24.8 25.9 25.6 25.3 24.9 24.8 24.5 24.3 23.0 20.9
Relative humidity, %
Mean max. 90.6 92.2 91.6 90.7 92.2 91.5 91.8 93.2 94.8 94.3 92.5 90.0
Mean 72.1 75.7 76.0 76.0 78.4 78.5 79.3 80.2 82.8 82.2 77.5 72.5
Mean min. 48.6 53.4 55.2 55.8 60.1 62.3 63.5 63.9 66.0 65.6 59.4 52.1
For AC case, the survey was conducted in classrooms and offices of six air-
conditioned buildings during February and early March. Student subjects were
interviewed after their classes. A total of 14 classes were surveyed. The survey was
conducted without any prior adjustment to indoor conditions. This means the subjects
were allowed to set the thermostats that adjust the temperature in the classrooms as
they normally did. A total of 229 responses were collected.
For NV case, the survey was conducted in a large student cafeteria from early to
late March when the weather was warming up as summer approached. Ceiling fans
were installed on the ceiling of the cafeteria. Students normally do not adjust the
speed of the fans, nor turn them off. A total of 271 responses were collected. Some
students participated in both cases.
2.2 Mehran Survey (NV class rooms)
Mehran University of Engineering and Technology is a government funded institute to
provide undergraduate degree education. It is located fifteen kilometers from
Hyderabad in Sindh Province, southern Pakistan. Hyderabad is located at latitude
25.4° N, just north of the Tropic of Cancer. The climate is hot for most times of the
year. A summary of the monthly dry-bulb temperatures appears in Table 2.
Table 2 Temperature profiles of Hyderabad, latitude 25.4 o N.
Temperature Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Mean max. 23 27 33 37 39 38 35 35 35 35 30 25
Mean 17 20 26 29 32 32 31 30 30 28 23 18
Mean min. 11 14 18 22 25 27 26 26 25 22 18 13
The subjects for the survey comprised young males only. In the survey, each student
was given a sheet of questionnaire at the end of a class in NV classroom. Students
were seated during a class. The survey was conducted during July and August, in the
period of hot and humid weather. A total of 764 responses were collected from 647
respondents. Thirty-four survey sessions were conducted in seven NV buildings. It
was noted that even though males increasingly wore jeans, traditional cultural practice
required even males to wear long trousers and discouraged them from wearing shorts
or light clothing.
3. RESULTS OF THE ASSESSMENT
Responses from respondents on personal variables of clothing and activities were used
to compute insulation values (clo) and metabolic rate (W/m2) of the respondents.
Measured values of environmental variables were used to compute operative and
effective temperatures, and predicted mean vote (PMV) in accordance with the
procedure in ISO 7730. The results are used with thermal sensation responses from
the subjects in the analyses that follow.
3.1 Summary of personal parameters and environmental condition
The values of personal variables and environmental conditions in the two locations
AIT Campus We assume that the activity level of all the subjects involved in the
study was at 1.2 met. The atmosphere in AIT is usually relaxed. Both male and
female students wore shorts and short-sleeved shirts, trousers, T-shirt, jeans, sandals
and shoes for both AC and NV cases. The relatively low clothing value of 0.22 clo in
Table 3 reflects the casual dress code of students around the campus and a clo value
of 0.26 reflects dressing style of European women in the class. Generally women
wore more garments than men and the maximum values of 1.02 and 1.04 are from
women in both AC and NV cases. These cases corresponds to a subjects wearing light
jackets in the classes. Temperatures in the AC classrooms were rather low. In NV
spaces, temperature and relative humidity values are relatively low because the
periods of survey were in cool and dry season, although hot and dry season was
approaching. The maximum value of air speed in NV space is not normally
Table 3 Summary of personal and environmental values in AIT campus.
Ta RH, % Tr Air Clothing Effective Age
Statistics (Air) (air) (Radiant) Velocity temperature
(oC) (kg/Kg) (oC) (m/s) (clo) (oC) Years
Cases AC NV AC NV AC NV AC NV AC NV AC NV AC NV
Mean 22.8 29.4 56.3 63 23.1 29.2 0.13 1.79 0.58 0.49 23.3 30.3 <=30 <=30
Std. dev. 1.5 1.4 7.6 8.6 1.6 1.4 0.08 2.33 0.14 0.13 1.7 1.0 157 204
Max 27.5 31.9 78.8 83.5 26.3 31.7 0.3 9.26 1.04 1.02 27.8 32.9 >30 >30
Min. 20.9 25.5 46.3 48.7 21 25.6 0.02 0.04 0.26 0.22 21.0 26.4 70 64
Note. Figures in the last column show number of subjects in each age group.
Mehran Unversity Table 4 shows that clothing insulation, clo value, varies
from as low as 0.51 to as high as 1.12 with an average of 0.82. These values are
higher than those found in other studies. For example, the average clo values in Table
3 for AIT campus and that found by Busch (1992) for respondents from Bangkok are
around 0.5, and that of Kwok (1998) for respondents from Hawaii was 0.36. Due to
the traditional and cultural norms students were discouraged from wearing shorts or
other light clothing. Given the current trend many students were found wearing jeans
and shoes. The observed indoor air temperature ranges from 30.9oC to as high as
38.2oC during the period of study. The observed indoor air velocity ranged from as
low as 0.2 m/s to as high as 1.42 m/s, and averaged to 0.74 m/s. The use of ceiling
fans was very common in NV classrooms of the university. The metabolic rate was
determined from the activities given by the subjects. It ranges from 55 W/m2 to 67.25
W/m2. Relative humidity were around 50%. None of the conditions in the classrooms
was found to fall within the ASHRAE comfort zone.
Table 4 Summary of personal and environmental values in Mehran.
Ta RH, % Tr Air Clothing Met. Rate Operative Effective
Statistics (Air) (air) (Radiant) Velocity temperature temperature
(oC) (kg/Kg) (oC) (m/s) (clo) (W/m2) (oC) (oC)
Average 34.55 48.02 34.15 0.81 0.82 61.13 34.4 34.2
Std. dev. 1.75 6.54 1.77 0.32 0.11 1.81 1.76 1.55
Min. 30.90 33.46 30.30 0.20 0.51 55.00 30.5 29.4
Max. 38.20 62.57 38.00 1.42 1.12 67.25 38.1 40.1
3.2 Thermal acceptability
ASHRAE Standard 55-2004 classifies an environment as thermally acceptable when
80 % of occupants are satisfied. This means that an environment receiving thermal
sensation vote from -1 to 1 is deemed thermally acceptable. Such environment should
encompass those that receive ‘neutral vote’ and those that receive ‘no change’ vote of
McIntyre scale. In our presentation, effective temperature is used, since relative
humidity of the environments changes during the course of survey, particularly for
AIT Campus We consider first the AC case. Figure 1 shows plots of ‘neutral
votes’, ‘acceptable votes’, and ‘no-change votes’, all expressed as percentages of the
total votes in each temperature bin of 0.5C effective temperature.
Neutral, acceptable, and no chage votes- Dis s atis fied and dis comfort votes -
AC case %Neutral AC cas e
60 %No change
40 Poly. (%
(%Acceptable 20 Discomfort)
Poly. (%No 10
0 change) (%Dissatisfied)
20 22 24 26 28
Poly. 20 22 24 26 28
ET* (%Neutral) ET*
a) Neutral, acceptable and no-change b) Dissatisfied and discomfort
Figure 1. Graphs of thermal acceptability and of persons dissatisfied at given
temperature bins for AC case.
It seems that the number of subjects may not be sufficient to render the graphs clearer,
or to allow better trend lines to be fitted. However, sufficiently clear trend lines can
be fitted for the case of neutral and acceptable votes in Figure 1a). The two peaks of
the trend lines for acceptable votes and neutral votes exhibit preferences for the
environment near effective temperature of 24 to 25 C and are very close to each other.
Alternatively, the trend line for dissatisfied votes-this is opposite to acceptable votes-
also shows a clear pattern of minimum dissatisfaction at around 25 C in Figure 1b).
The trend line for discomfort votes-as opposite to ’no-change’ votes is not as clear.
These results illustrate that the 7-point scale of ASHRAE enables more conclusive
information to be drawn than the McIntyre scale.
Figure 2 illustrates results for the NV case. The trend lines formed from ASHRAE
sensation votes also appear more indicative and more informative than those resulting
from the McIntyre results.
Neutral, acceptable, and No change votes - Dissatisfied and discom fort votes-NV case
NV cas e
100 %Neutral 100 %Discomfort
80 %Accept 80
60 %No change 60
40 Poly. 40
20 Poly. 20
0 0 Poly.
Poly. (%No (%Dissatisfied)
26 28 30 32 34 change) 26 28 30 32 34
a) Neutral, acceptable and no-change b) Dissatisfied and discomfort
Figure 2. Graphs of thermal acceptability and of persons dissatisfied at given
temperature bins for NV case.
The trend lines in Figure 2 imply that the neutral environment corresponds to a
temperature of over 28 C.
Mehran Unversity The number of subjects in this case is larger and the overall
results are clearer. The graphs for neutral and no-change votes in Figure 3a) are
almost identical and both imply that neutral temperature is further to the left of the
plot area. The pattern of pattern of acceptable votes is also clear.
Neutral, No change , and Acceptable Dis s atis fied and dis comfort votes
100 100.0 Dissatsfied
80 %Neutral 80.0
60 %No change 60.0
40 40.0 (Discomfort)
20 20.0 Poly.
30 32 34 36 38 30 32 34 36 38
Poly. (%No ET*
a) Neutral, no-change, and acceptable b) Dissatisfied and discomfort
Figure 3 Graphs of thermal acceptability and of persons dissatisfied at given
temperature bins for NV condition at Mehran University.
The graph of dissatisfied votes in Figure 3b), as opposite to acceptable votes, also
shows a clear pattern, while the pattern of discomfort votes, as opposite to no-change
votes, also becomes clear. The pattern of discomfort votes seem to imply that neutral
temperature would be much lower than that can be predicted from the plot, while the
graph of dissatisfied votes imply that neutral temperature is near the left end of the
graph. The fact that it was very warm during the period of the survey appears to be
clear from the results above.
3.2 Mean thermal sensation
The results of indirect thermal acceptability votes, or the more direct votes of
McIntyre, when graphed show certain pattern on acceptable or unacceptable
environment, but do not seem to yield a clear point of neutrality. Analysis of mean
thermal sensation, on the other hand, will lead to directly identifying a single neutral
or most acceptable environment.
AIT Campus We first consider the AC case. Figure 4 shows plots of weighted
thermal sensation votes (WTSV), based on ASRAE scale, and the regressed linear
y = 0.2267x - 5.7201
Weighted TSV-AC case
R2 = 0.8737
20 22 24 26 28
Effective Temperature (ET*)
Figure 4 Plots of weighted thermal sensation vote and regressed line.
The regressed neutral effective temperature is obtained as TN = 25.23. This is higher
than that of Busch (1992) of 24.5 C. The value of the coefficient of ET* of 0.2267
implies that the neutral temperature is not sensitive to temperature change.
Weighted TSV y = 0.505x - 12.21 Weighted TSV y = 0.3504x - 8.6479
for Male-AC Case R2 = 0.8338 for Fem ale-AC case R2 = 0.8453
TSV (Male) Linear (TSV (Male)) TSV(Female) Linear (TSV(Female))
20.00 22.00 24.00 26.00 28.00 20 22 24 26 28
Effective temperature (ET*) Effective temperature (ET*)
Figure 5 Plots of WTSV and regressed line, male and female cases.
Figure 5 shows the results of regression of WTSV for the case of male and female
subjects. The neutral temperatures obtained are 24.2 C and 24.7 C respectively. Even
though the number of subjects in each case is about half of the total for AC case, the
resultant values of coefficient of determination are high in both cases, which show
that TSV obtained are consistent. The results do not show any significant difference
between the male and female subjects.
Figure 6 shows the results of regression of WTSV for the group of subjects with
ages less than or equal to thirty years old and for the group of subjects with ages
exceeding thirty. The neutral temperatures obtained are 25.4 C and 24.5 C
respectively. The difference of almost one degree can be significant, but the
coefficient of determination of the regression in the higher age group is low,
signifying some uncertainty of the results.
Weighted TSV y = 0.2274x - 5.7802 Weighted TSV y = 0.3336x - 8.1761
for Age<=30-AC case R2 = 0.8526 for Age>30)-AC case R2 = 0.6538
TSV(Age<=30)) Linear (TSV(Age<=30))) TSV(Age>30) Linear (TSV(Age>30))
20 21 22 23 24 25 26 27 28 0.5
-0.5 20 21 22 23 24 25 26 27 28 29
Effective temperature (ET*) Effective temperature (ET*)
Figure 6 Plots of WTSV and regressed line, cases of younger and older subjects.
We next consider the NV case. Figure 7 show plots of WTSV and regressed line for
subjects in NV environment.
Weighted TSV-NV case y = 0.3911x - 11.034
R2 = 0.8245
ASHRAE Linear (ASHRAE)
-0.5 26 28 30 32 34
Effective temperature ET*
Figure 7 Plots of WTSV and regressed line, NV case.
The neutral temperature is obtained as TN = 28.17 C. The number of subjects was
sufficiently large to render a reasonable value of coefficient of determination.
When the value of neutral temperature above is compared to that of Busch (1992)
of 28.5 C, the two values are close and are significantly different from the neutral
temperature obtained for AC environment above. The case here differs from that of
Busch. The subjects in the AC environment and the subjects in the NV environment in
the case of Busch are from different groups of people, but for our case here most of the
subjects in the two environments are the same persons. It is not the case of
acclimatization that one person feels comfortable in two different environments a
Figure 8 shows results male and female, and Figure 9 shows results for the two
age groups. In some of these cases, the values of the coefficient of determination are
Weighted TSV y = 0.3718x - 9.9167 Weighted TSV y = 0.418x - 11.299
for Male-NV case R2 = 0.71 for Fem ale-NV case R2 = 0.594
TSV( male) Linear (TSV( male)) TSV(female) Linear (TSV(female))
1 Thermal sensation 1
0.5 value (TSV) 0.5
-0.5 24 25 26 27 28 29 30 31 32 -0.5 24 25 26 27 28 29 30 31 32
Effective temperature (ET*) Effective temperature (ET*)
Figure 8 Plots of WTSV and regressed line, male and female cases, NV environment.
Weighted TSV y = 0.3083x - 8.4964 Weighted TSV y = 0.4176x - 11.932
for Age<=30)-NV case R2 = 0.8457 for Age>30-NVcase R2 = 0.5876
Linear (TSV(Age<=30)) Linear (TSV(Age>30))
1 26 27 28 29 30 31 32 33 34
26 27 28 29 30 31 32 33 34 -3
Effective temperature (ET*) Effective temperature (ET*)
Figure 9 Plots of WTSV and regressed line, cases of younger and older subjects, NV
The values of neutral temperature for male and female are 26.7 and 27.0 C
respectively. For the case of the younger age group, the value is 27.6 C, and that of
the older age group, the value is 28.6 C. The number of subjects in the younger age
group is larger than that of the older group. The temperature difference is not
insignificant. However, here the older age group preferred higher temperature in NV
environment, while the same group preferred lower temperature in AC environment.
There are results for subjects of the same nationality, but the number of subjects in
each group is too small.
Mehran Unversity The subjects comprised young male only. Figure 10 show
plots of weighted TSV and the corresponding regression line.
Weighted TSV y = 0.2955x - 8.8196
R2 = 0.8701
TSV Linear (TSV)
25 30 35 40
Figure 10 Plots of weighted TSV and the regression line.
The neutral temperature is obtained in this case as TN = 29.85 C. This temperature is
significantly higher than the value obtained for NV space in AIT campus. This
reflects the prevailing thermal environment during the time of survey.
3.3 Predicted mean vote
We consider only the NV case. Figure 11 shows the relationship between TSV and
PMV from both AIT and Mehran surveys.
y = 1.2533x - 0.8765 y = 0.7845x - 1.0477
TSV and PMV TSV and PMV
2 R2 = 0.8723
NV case, AIT R = 0.8288 Mehran
0 1 2 3 0 1 2 3 4 5
a) AIT b) Mehran
Figure 11 Relationship between TSV and PMV.
Both results show that actual thermal sensation vote are lower than PMV from
ASHRAE Standard 55-2004. For the case of AIT, the line crosses the horizontal axis
at +0.779, while for Mehran it crosses at +1.34. The result for Thailand is similar to
that of Busch (1992), at +0.5. The value obtained for Mehran is much higher than the
Thailand case and is also higher than the result of Henry and Wong (2004) at +1.2 for
3.4 Probit analysis
Probit analysis is to results from sensation votes using McIntyre scale, both for warm
votes and cooler votes, each case at a time.
AIT Campus Figure 12 shows results of probit analysis for AC environment and
NV environment from the survey in AIT.
Probit Analysis for NV Probit Analysis for AC
Cooloer Warmer Cooler Warmer
161820222426283032343638 12 1416 18 20 2224 26 2830 32
Figure 12 Probit analysis results from AIT survey.
The results indicate that the temperature where the warmer line intersects the cooler
line at 24 C for the AC case and at 26.7 for the NV case. These results differ from
those obtained from analysis of mean thermal sensation.
Mehran University Figure 13 shows results of probit analysis for the results of
Mehran survey. The unbalanced environmental condition, where high temperature
prevailed throughout the period leads to an effective temperature of 34 C at the
30.0 35.0 40.0
Figure 13 Results of probit analysis for the Mehran survey.
The result of probit analysis for the case of Mehran University shows a short coming
of the three-point McIntrye scale where the final result is heavily biased towards
higher temperature because of lack of data at lower temperature. The lack of
numerical score for thermal sensation may also render numerical result less accurate.
3.5 Adaptive models
Researchers conducting field study of thermal comfort have proposed simple models
that relate neutral temperature to prevailing air temperature. Table 5 shows some
models chosen. These models have been proposed based on the premise that neutral
temperature in a given environment is related to the prevailing outdoor temperature of
the location, To, generally taken as the mean outdoor air temperature around the given
month. Some model relates neutral temperature to prevailing indoor temperature, Ti,
also taken as the mean indoor air temperature. Some models utilize both indoor and
Table 5 Some adaptive models for predicting neutral temperature.
Type of space Model Reference
NV Tn = 17 +0.38To Nicol and Roaf (1996)
NV Tn,i = 2.6 + 0.831Ti Humphreys (1976)
NV Tn,o = 11.9 + 0.534To Humphreys (1976)
AC 24* 2
Tn =23.9+0.295(To -22)e Humphreys (1978)
NV Tn,i = 5.41 +0.73 Ti Auliciems and deDear (1986)
NV Tn,o = 17.6 + 0.31To Auliciems and deDear (1986)
NV Tn,i,o = 9.22 +0.48Ti +0.14To Auliciems and deDear (1986)
A comparison of the neutral effective temperatures, ETN , obtained in the studies at
AIT and Mehran with values using the above models are shown in Table 6.
Table 6 Neutral temperatures calculated from adaptive models.
Type of Value
Campus space To Ti ETN Model from
AIT NV 29.6 29.4 28.17 Nicol and Roaf (1996) 28.2
AIT NV 29.6 29.4 28.17 Humphreys (1976) 27.0
AIT NV 29.6 29.4 28.17 Humphreys (1976) 27.7
AIT AC 29.6 29.4 25.23 Humphreys (1978) 26.3
AIT NV 29.6 29.4 28.17 Auliciems and deDear (1986) 26.9
AIT NV 29.6 29.4 28.17 Auliciems and deDear (1986) 26.8
AIT NV 29.6 29.4 28.17 Auliciems and deDear (1986) 27.2
Mehran NV 32.5 33.7 29.85 Nicol and Roaf (1996) 29.4
Mehran NV 32.5 33.7 29.85 Humphreys (1976) 30.6
Mehran NV 32.5 33.7 29.85 Humphreys (1976) 29.3
Mehran NV 32.5 33.7 29.85 Auliciems and deDear (1986) 30.0
Mehran NV 32.5 33.7 29.85 Auliciems and deDear (1986) 27.7
Mehran NV 32.5 33.7 29.85 Auliciems and deDear (1986) 29.9
It seems that for AIT case, adaptive models under estimates NV effective temperature,
except for Nicol and Roal model. Humphreys’ model for AC case predicts reasonably
well. For Mehran case, all models perform well
4. RADIANT COOLING FOR THERMAL COMFORT
In hot and humid climate, natural ventilation and stack ventilation are often mentioned
as means to achieve thermal comfort.
700 100 Iglobal
60 Tamb (oC)
200 Tsky (oC)
0 4 8 12 16 20 24
Figure 14 Weather data of a typical day near the end of rainy season in Bangkok.
When we examine the profiles of some meteorological variables such as those in
Figure 14 for a typical day near the end of the rainy season for Bangkok, one should
realize how viable these options are. Wind speeds are low in most locations where
temperatures are high, typically averaged to 1-2 m/s. This renders natural ventilation
difficult. From a nighttime low of 25 C, the air temperature reaches 35 C during
daytime. To use night air for cooling is not practical. To create stack effect for air to
flow through the interior through stack, the air in the stack must be sufficiently higher
than 35 C and the entry of the stack must be near ground level to create sufficient air
flow. These simple natural means do not seem to be viable.
When Baruch Givoni (2000) was invited to present his view on passive
architecture concept for hot humid climate, he suggested use of natural ventilation,
night air cooling, cooling by thermal radiation of surfaces to sky, indirect evaporative
cooling, and utilization of soil as a cooling source. The first two options are not
viable as discussed above. If we examine the sky temperature in Figure 14, we see
that it is very close to air temperature. This is true for most days of a year, except in
cool season. This implies that cooling by thermal radiation to sky is also not viable.
Records from the Meteorological Department on soil temperature under the surface of
ground (up to one meter) show that underground temperature is close to long-term
average air temperature. So utilization of soil as a cooling source may not be viable.
This leaves only indirect evaporative option.
Radiant cooling by passing cool water through panels on ceiling and walls may be
a viable means of application of indirect evaporative cooling. Earlier studies reported
applications of radiant cooling in Europe during summer, Mumma (2001). Cooling
water is supplied to the panels typically at 18-20 C. This water could be obtained
directly from tap water or from refrigeration. It has also been suggested that cooling
tower could be used to regenerate the cooling water, Facao and Oliveira (2000). In
the latter case the temperature of supply water would be related to the wet-bulb
temperature of the ambient air. The temperature difference between the supply cooling
water and the ambient air could be limited. This will lower the cooling capacity of the
cooling panel and necessitate increasing the size of the cooling panel to meet load.
Nevertheless, radiant cooling system based on use of cooling tower for regeneration
of cooling water would be suitable for application in cases where cooling loads are
relatively low. Table 7 exhibits the number of hours that dew-point temperature and
wet-bulb temperature exceed given levels in a year. This illustrates the extent that
radiant cooling could be used, and the capacity it would have per unit area of the
panel when cooling tower is used to regenerate cooling water.
Table 7 Number of hours in a year that wet-bulb temperature and dew-point
temperature exceed given values.
Wet-bulb temperature Dew-point temperature
Number of hours Number of hours
Day Night Level Day Night
Bangkok Hyderbad Bangkok exceeded Bangkok Hyderabad Bangkok
25.0 1774 NA 333 25.0 49 NA 34
26.0 372 425 5 25.5 5 NA 0
27.0 7 252 0 26.0 1 133 0
27.5 1 157 0 27.0 0 52 0
28.0 0 111 0 27.5 0 25 0
28.0 0 10 0
Experimental and simulation studies on application of radiant cooling for comfort
conditioning has been undertaken at the Asian Institute of Technology (AIT) for some
time. Our experimental study on radiant cooling at AIT was constrained by the limited
capacity of the system, but simulation study using TRNSYS simulation program
showed that radiant cooling using cooling tower for regeneration of cooling water was
viable, Vangtook and Chirarattananon (2006), Vangtook and Chirarattananon (2005).
Supplementary means such as use of electric fan to increase air speed through human
bodies may be needed to achieve thermal comfort.
AIT Campus An experimental room equipped with a set of radiant panels of total
area of 7.5 m2 was used for physical experiment. Figure 15 shows its configuration.
placed to Insulation Plenum
shade solar East
windows sections of
Cooling panel walls
a) A photograph of the experimental b) A diagram of the interior of the
room (without the shading board.) experimental room showing the
locations of the cooling panel.
Figure 15 A photograph and a diagram of the experimental room.
The capacity of the panel was insufficient to cope with large load, so experiments
were conducted during nighttime in hot season, and all day during cooler season. The
temperature of cooling water was kept mainly to 24-25 C to avoid condensation of
moisture from air on the panel. Figure 16 exhibits heat flux or cooling extracted by
Radiation Total Radiation Total
Heat flux, ceiling panel(W/m2)
Heat flux, wall panel (W/m2)
Time (hh:mm) Time (hh:mm)
a) Ceiling panel b) Wall panel
Figure 16 Graphs of total heat fluxes and radiation heat fluxes received by the panels.
The size of heat flux of 30-40 Wm-2 limits the capacity of the set to 200-350 W even
when refrigeration was used to cool the water that was supplied to the panel.
However, when the load is low, the set up was able to reduce temperature and
maintain condition in the room to acceptable level as seen from the graphs of PMV in
Computer simulation with TRNSYS was then used to further investigate to see the
size of radiant panel required and if cooling tower could be used to regenerate cooling
water. The results show that a radiant panel of 26 m2 supplied with water from cooling
tower could be used to provide cooling to a room the same size of the experimental
room of 16 m2. The environment in the room could be kept to neutral level for 86 %
and to acceptable level for 100% of all hours in a year for sedentary activity, provided
electric fan is used as supplement. If the air could be kept dryer, electric fan is not
2 PMVm PMVm
1.5 PMVT2 1 PMVT2
Time (hh:mm) Time (hh:mm)
(a) first night (b) second night
Figure 17 Calculated PMV from measured values and from TRNSYS simulation,
Some experimental work on solar-regenerated desiccant dehumidification has also
been undertaken. Desiccant dehumidification could be used to remove latent load that
radiant panel is incapable of.
Mehran University A simulation study on the application of radiant cooling in
classrooms using cooling tower for regeneration of cooling water was conducted. As
can be surmised from the relatively wet-bulb temperature of Hyderabad in Table 7,
supplementary fans would have to be used to keep the internal environment at
acceptable level. The simulation shows that maximum temperature in the interior
reached 34 C. Nevertheless, radiant cooling system is capable of modulating the
interior environment to satisfactory level. Figure 18 illustrates typical interior
0 24 48 72 96 120 144 168
AIR DRY BULB OT
Figure 18. Interior air and operative temperatures (OT) and dry-bulb temperature of
the exterior air when radiant cooling is used for the classroom in April.
The fact that people could feel comfortable a moment apart in two distinct
environments warrants further investigation. In hot and humid climate, substantial
energy could be saved and energy sustainability could be achieved if thermal comfort
could be attained at higher temperature. This paper shows some potential of
application of passive cooling, but this should be further explored.
• ASHRAE. Thermal Environmental Conditions for Human Occupancy.
ANSI/ASHRAE Standard 55-2004.
• Auliciems, A.,and deDear, R.. 1986. Air-conditioning in a tropical climate: Impacts
upon European residents in Darwin, Australia. International Journal of
Biometeorology, 30(3), 259-282.
• Busch, J.F. 1992. A tale of two populations: Thermal comfort in Air-conditioned
and naturally ventilated offices in Thailand. Energy and Buildings, 18(3), 235-249.
• EN ISO 7730. 1994. Moderate thermal environments-determination of the PMV and
PPD indices and specification of the conditions for thermal comfort, ISO Geneva.
• Facao, J. and Oliveira, A.C. 2000. Thermal behavior closed of wet cooling towers
for use with chilled ceilings, Applied Thermal Engineering, 20, 1225-1236.
• Givoni, Baruch. 2000. Building Design and Passive Cooling for Hot Humid Region,
Proceedings of a conference on Green Architecture: the Sustainable Built
Environment in the New Millennium, Bangkok, April 21-22.
• Henry F. and Wong N. H. 2004. Thermal Comfort for naturally ventilated houses in
Indonesia, Energy and Buildings, 36(7), 614-626.
• Humphreys, M.A.. Comfortable indoor temperatures related to the outdoor air
temperature. BRE Publication Draft. Garston: Building Research Establishment,
• Humphreys, M.A.1978. Building Research Practice, 6 (2), 1.
• Jitkhajornwanich, K. and Pitts, A. 2002. Interpretations of thermal reponses of four
subject groups in transitional spaces of buildings in Bangkok, Building and
Environment, 37, 1193-1204.
• Khedari, J., Yamtraipat, N., and Hirunlabh, J. A new concept for setting thermal
comfort standards, Proceedings of the conference on Moving thermal comfort
standards into the 21st century, Cumberland Lodge, Windsor, UK, 5-8 April 2001.
• Kwok A.G.. 1998. Thermal Comfort in Tropical Classrooms, ASHRAE Transactions
• Lomonaco, C. and Miller, D. 1997. Comfort and Control in the Workplace.
ASHRAE Journal, September Issue.
• Malama, A., Sharples, S., Pitts, A.C., and Jitkhajornwanich, K. 1998. An
investigation of the thermal comfort adaptive model in a tropical upland climate,
ASHRAE Transactions 1A, 1194-1203.
• Mumma, S.A. 2001. Ceiling panel cooling systems, ASHRAE Journal, November,
• Nicol, J.F. and Humphreys, M.A. 2002. Adaptive thermal comfort and sustainable
thermal standards for buildings, Energy and Buildings, 34, 563-572.
• Nicol, F. and Roaf, S. Pioneering New Indoor Temperature Standards: the Pakistan
Project. Energy and Buildings, 1996, 23, 169-174
• Vangtook, P. and Chirarattananon, S. 2006. An Experimental Investigation of
Application of Radiant Cooling in Hot Humid Climate, Energy and Buildings, 38,
• Vangtook, P. and Chirarattananon, S. 2005. Application of Radiant Cooling as a
Passive Cooling Option in Hot Humid Climate, Building and Environment, in press
and available online.
• Wong, N.H. and Khoo, S.S. 2003. Thermal comfort in classrooms in the tropics,
Energy and Buildings, 35, 337-351.