Full text of DOI:10.1080/10789669.2002.10391436

This article was downloaded by: [Chulalongkorn University]
On: 26 December 2014, At: 20:05
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
HVAC&R Research
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/uhvc20
Experimental Study of Factors That Affect Thermal
Comfort in an Upward-Displacement Air-Conditioned
Room
Zhiwei Lian a & Haiying Wang b
a Institute of Refrigeration and Cryogenics in the School of Power and Energy Engineering at
Shanghai Jiaotong University , China
b Qingdao Institute of Architecture and Engineering , China
Published online: 28 Feb 2011.
To cite this article: Zhiwei Lian & Haiying Wang (2002) Experimental Study of Factors That Affect Thermal Comfort in an
Upward-Displacement Air-Conditioned Room, HVAC&R Research, 8:2, 191-200
To link to this article: http://dx.doi.org/10.1080/10789669.2002.10391436
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

VOL. 8, NO. 2
HVAC&R RESEARCH
APRIL 2002
Experimental Study of Factors That Affect
Thermal Comfort in an Upward-Displacement
Air-Conditioned Room
Zhiwei Lian, Ph.D.
Haiying Wang
Experiments were carried out to determine the effect of an underfloor air-conditioning system on
comfort. The four factors evaluated were the type of outlet, the distance between the outlet and the
occupant, the velocity of the supply air, and the temperature of supply air. It was found that the
distance between occupant and outlet has a significant influence on thermal comfort, the velocity
and temperature of the supply air have a moderate influence, and the type of outlet has little influ-
ence. Recommended values of these factors were developed that would achieve thermal comfort.
The results provide reference data on air distribution in underfloor air-conditioned rooms.
INTRODUCTION
It is well known that comfortable indoor thermal environments improve work efficiency and
benefit health. Further, maintaining suitable indoor environmental conditions has the potential to
reduce energy consumption in buildings. An upward-displacement air-conditioning system can
provide better indoor air quality (Lian and Ma 1996) and can save building energy. Many coun-
tries have been engaged in research into these systems. During the 1980s, the upward-displace-
ment air-conditioning systems of Lloyd’s Building in London and Huifeng Bank in Hong Kong
were successfully designed, and attracted attention in the air-conditioning field. In the following
decade, nearly 20 air-conditioning equipment corporations and universities in Japan set up spe-
cialized laboratories to study upward-displacement air-conditioning systems (Fan 1997).
However, disagreements about the thermal comfort of upward airflow still exist. The adop-
tion of a system with upward airflow requires the direct supply of treated air to the occupied
zone, which leads to lower temperatures at the bottom and higher temperature at the top of
the space. In addition, the outlets installed at the bottom are near the occupants and require
a shorter attenuation distance. If not properly designed, the flow from these outlets will cause
discomfort. This problem has been studied widely; for example, Sodec and Craig (1990) have
done research based on the thermal comfort theory of Fanger (1967) and have summarized the
European experience with upward airflow. They established that if upward airflow systems were
designed according to certain standards, a comfortable environment could be achieved. Japanese
researchers have used a copper manikin to simulate the Equivalent Homogeneous Temperature
(EHT) of different areas of the human body. They concluded that if the difference in EHT
between the head and feet is less than 3°C, the environment would satisfy comfort requirements
(Fan 1993). Other researchers also have studied this subject and made recommendations (Fan
1993; Lian and Feng 2002; Ma 1997). However, a method for guiding the design of air around
practical projects in upward-airflow systems does not exist. Research studies (Fan 1993; Fanger
1967; Int-Hout 1990; Hanzawa and Nagasawa 1990; Ma 1995) have shown that thermal comfort
is the result of many physical factors along with the mental, subjective psychological reaction.
Zhiwei Lian is a professor at the Institute of Refrigeration and Cryogenics in the School of Power and Energy Engineer-
ing at Shanghai Jiaotong University, China. Haiying Wang is a lecturer at the Qingdao Institute of Architecture and En-
gineering, China.
191

192
HVAC&R RESEARCH
Studies in the field are needed to make clear the contribution of different factors that affect the
body comfort in upward airflow and to recommend design values.
FACTORS INFLUENCING THERMAL COMFORT
In an upward-displacement system, treated air is supplied in the horizontal or vertical direc-
tion through outlets located in the lower part of the zone. The fresh air passes through the occu-
pant zone first and then rises up through the unoccupied zone to the exhaust. Thus, the
concentrations of humidity and pollutants and the temperature of the exhaust air are higher than
those in the occupant zone. In this manner, the system provides a better environment and
reduces air-conditioning energy. The mode of regulation is flexible so that these systems are
adaptable to different types of buildings and air-distribution systems. These systems have poten-
tial for modern office buildings.
Previous research into upward-airflow systems has shown that the main indoor environment
indexes that affect body comfort are mean temperature in the occupant zone, humidity, air
velocity, and mean radiation temperature. Mean temperature and air velocity in the occupant
zone are closely related to air distribution. The type of outlet, the location and number of outlets,
and the velocity and temperature of supply air are factors that influence the two indexes of mean
temperature and velocity.
The influence of the type of outlet on the distribution of air is related to its structure and per-
formance. In the active region of the outlet, different outlet types lead to different distribution pat-
terns. The layout of outlets is important. When the supply air outlets are located in the floor, they
need to be some distance away from the occupant to prevent drafts, which can result in discomfort.
Supply air velocity is an important design variable for air distribution. With the same volume
of air, a higher supply velocity means less supply air area and smaller number or size of the out-
lets. However, a higher velocity leads to a longer attenuation distance and the danger of blasting
air on the occupants.
Another important factor is temperature of supply air. A lower temperature of the supply air
produces a larger air temperature difference for heat transfer. A lower volume of supply air is
required, which reduces the plant capacity and fan power. However, it leads to large temperature
gradients in the zone and irregular air distribution, which is worse for upward airflow. On the
other hand, increasing the temperature of supply air will increase the required volume of supply
air, air velocity, and size and number of outlets.
The goal of this study is to evaluate the effects of the following four factors: (1) type of outlet,
(2) distance between occupant and outlet, (3) velocity of the supply air, and (4) temperature of
the supply air.
EXPERIMENTAL SCHEME
Three types of outlets with the same diameter (200 mm) were used in this study: a circular
straight vane outlet, a rotational flow outlet, and an inductive rotational flow outlet. The circular
straight vane outlet has a simple structure with little flow resistance but poor airflow diffusivity.
In the rotational flow outlet, air enters at the bottom and flows upward against vanes that pro-
duce a rotating flow. This outlet has better diffusivity than the circular outlet, but a more com-
plex structure and larger flow resistance. The inductive rotational flow outlet induces air from
the air-conditioned room, with the result that the temperature of air coming from the outlet is
higher than with the former types.
Three levels of distance between occupant and outlet were studied: 0.4, 0.7, and 1.0 m. Some
previous research concluded that when the distance between occupants and the edge of outlet
was kept larger than 0.4 to 0.6 m, the predicted percentage dissatisfied PPD because of drafts
will be less than 15% (Wang 1986). Other research suggested that the distance should be kept

VOLUME 8, NUMBER 2, APRIL 2002
193
Table 1. Orthogonal Experiment Scheme
Factor
Experiment
Code
C: Supply Air
Velocity, m/s
D: Supply Air
Temperature, °C
A: Outlet Type
1: Circular straight vane
1: Circular straight vane
1: Circular straight vane
2: Rotational flow
B: Distance, m
1: 0.4
1
2
3
4
5
6
7
8
9
1: 0.5
2: 1.0
3: 1.5
2: 1.0
3: 1.5
1: 0.5
3: 1.5
1: 0.5
2: 1.0
1: 18
2: 0.7
2: 20
3: 1.0
3:22
1: 0.4
3:22
2: Rotational flow
2: 0.7
1: 18
2: Rotational flow
3: 1.0
2: 20
3: DY inductive rotational flow
3: DY inductive rotational flow
3: DY inductive rotational flow
1: 0.4
2: 20
2: 0.7
3:22
3: 1.0
1: 18
between 0.8 and 1.0 m; other studies show that when the number of outlets is between 0.4 and
0.7 per square metre, the ventilation efficiency ET is greater than 2.0 and will increase as the
number of outlets is increased (Ma 1997). The selection of the three levels of distance was based
on these studies.
The range of velocity of supply air is between 2 and 5 m/s in conventional air distribution sys-
tems. This range is inapplicable to upward-airflow systems in which the velocity is generally
less than 2 m/s. The circular straight vane outlet has a maximum velocity of 2 m/s, the rotational
flow outlet is less, and the DY inductive rotational flow outlet has a maximum velocity of only
1.6 m/s. Considering the requirements for upward air systems and the performance of the out-
lets, three velocity levels of 1.5, 1, and 0.5 m/s were selected for study.
The temperature levels selected for study were 18, 20, and 22°C. These are consistent with
the requirements of upward-displacement air-conditioning systems.
This study used an orthogonal experimental design based on the number and the levels of fac-
tors in accordance with mathematical statistics. There are two tables that can arrange for testing
four factors at three levels: L934 and L27313. Each contains an interaction column. In the L934
table, 9 groups of experiments are needed, whereas in the L27313 table 27 groups of experiments
are required. The L934 table was selected and the 9 groups of experiments are shown in Table 1.
Physiological experiments on humans were carried out to determine the influence of the iden-
tified factors on body comfort. There are two methods currently used to evaluate the response of
humans to the thermal environment. One method is called a subjective evaluation method. The
subjects fill out tables that describe their own sensation about the environment. Usually a scale
of seven numbers ranging from –3 to +3, representing sensations of cold, cool, slightly cool,
neutral, lightly warm, warm, and hot, is used. Because the reaction of each individual to the
same environment is somewhat different, a large number of subjects is needed. The other
approach is the objective evaluation method. The skin temperature of the subjects is measured.
The mean skin temperature of the body tsk is an effective objective index that changes with
ambient conditions. For dry skin, this temperature index depends mainly on the ambient temper-
ature and velocity. The skin temperature difference ∆tsk, which is the difference in temperature
between the chest and foot, is another important index. The effect on thermal comfort of the
upward airflow characteristics (draft, temperature gradient) is embodied by ∆tsk. The smaller the
value of ∆tsk, the more comfortable the environment.
Both methods were used in this study. The analysis of the experiments depends mainly on the
∆tsk data and is synthesized with the results of the subjective responses. A large number of sub-
jects was not needed based on preceding experimental studies (Lian and Feng 2002; Ma 1995).

194
HVAC&R RESEARCH
Experimental Facility
All experiments were carried out in an air-conditioned upward-displacement room of size
5.0 by 2.9 by 3.5 m that was located in a surrounding air-conditioned space. The room floor was
racked wood. The outlets were fitted into the floor, with the number changeable if necessary.
The ceiling was made of wood with a rectangular opening of 900 by 900 mm at center that
served as the return air outlet. The structure of the room is shown in Figure 1.
The ambient condition was kept unchanged during experiments. The test ambient conditions
were an air temperature in the occupant zone of 25 1°C, a relative humidity of 50 20%, and
a mean radiation temperature of 27 2°C. These conditions provide a comfortable underfloor
air-conditioning environment, according to previous research. Use of the same experimental
conditions reduces the experimental errors caused by different indoor situations.
The indoor heat load was simulated by hot plates that were controlled by voltage tuned trans-
formers, which can be regulated to set different experiment conditions. The volume of the
supply air was controlled by a variable-speed fan and the heat load was regulated for each exper-
iment in order to keep the room at the desired condition. The value of voltage and electric cur-
rent were recorded to calculate the head load. The heat load ranged from 67 to 198 W/m2 in the
experiments, which is similar to the working mode in summer of modern offices and computer
rooms. When the velocity of supply air is low and the temperature of supply air is high, the heat
load would be small, and vice versa. The volume of fresh air was regulated by a plug valve to
keep the rate of fresh air introduction at about 30%.
Different airflow distributions have different influences on thermal comfort, which is
reflected by measuring both the thermal sensations of subjects and their skin temperature differ-
ence. The thermal sensation is divided into seven levels, as discussed previously. In addition,
questions were asked about draft sensation at the knees and feet. The skin temperatures were
measured by thermocouples. Five points were selected to represent different areas of the body:
forehead, chest, hand, knee, and foot. The weighted mean skin temperature tsk was calculated in
the following manner:
tsk = 0.07tforehead + 0.5tchest + 0.1thand + 0.25tknee + 0.08tfoot
where ∆tsk is calculated by
(1)
∆tsk = tchest – tfoot
(2)
Figure 1. Room Structure Used in Experiment

VOLUME 8, NUMBER 2, APRIL 2002
195
Two subjects, a male and a female, were tested in each experiment. They were arranged sit-
ting before two tables placed in a row and were able to read, write, or talk, but they were not
allowed to walk during the test. This situation is similar to that in offices. The subjects were not
close to the four outlets. The distance between occupant and outlet is defined as the minimum
range between the subject and the edge of the nearest outlet. Because of the confined space the
subjects usually sat near one or two outlets. The outlets could not be moved but were covered as
necessary. In some experiments all 12 outlets were working together but this was not always the
case.
The subjects wore clothes of the same thermal resistance. The clothing ensemble included a
semi-slip, thin trousers, silk hose, and slippers with thermal resistance of 0.19, 0.26, 0.03, and
0.04 clo respectively. With another 0.05 for underclothing, the total thermal resistance of clothes
was 0.53 clo.
Eighteen subjects were tested altogether, ranging in age from 22 to 30; they were volunteers
in good health. No subject experienced all of the nine experimental conditions. The objective
test index, skin temperature tsk, was used because it differs very little among different people
under the same environment. Thus, consistent experimental data could be obtained and a large
number of subjects was unnecessary.
All the experiments were carried out during the afternoon or evening in September when
the outdoor air temperature was very high. Each experiment lasted for an hour. The skin tem-
peratures of the subjects and the indoor air temperature were measured every 15 minutes. The
subjects were asked to fill out forms that recorded their thermal comfort sensation every 30
minutes.
Data Analysis
The analysis of the results shows little difference between the responses of the male and
female subjects. The index ∆tsk of the female subjects is a little greater than that of male sub-
jects, which implies that females are more sensitive to the environmental variables. The experi-
mental index for females is taken to be Yi, which is the average value of ∆tsk for one woman
tested in one experiment. The index ∆tsk was recorded four times in each experiment, and the
value had a tendency to increase during the course of the experiment. The temperature gradient
in the room is the main reason for this phenomenon, because the subjective records show that
most of the environments were comfortable, although in some cases the subjects were troubled
by drafts.
Figure 2 depicts the variation of the factors determined from the analysis. In this figure ∆tsk is
an average value over all subjects, A1 is the average value of factor A when it is at level 1, A2
Figure 2. Variation of Factors

196
HVAC&R RESEARCH
Table 2. Variance Calculation Table
Factor
A
1
1
1
1
2
2
2
3
3
3
B
2
1
2
3
1
2
3
1
2
3
C
3
1
2
3
2
3
1
3
1
2
D
4
1
2
3
3
1
2
2
3
1
Experimental Index Yi*
Experiment Code
Female’s ∆tsk, °C
1
2
3
4
5
6
7
8
9
3.89
2.92
2.51
2.80
3.60
2.00
4.38
2.09
2.83
I2
II2
III2
P
86.86
70.56
86.49
243.91
81.30
0.18
122.54
74.13
53.88
250.55
83.52
2.40
63.68
73.10
110.04
246.82
82.27
1.15
106.50
86.49
54.76
247.75
82.58
1.46
G = 27.02
K = 81.12
Y = 3.00
Q
S
The meaning of each symbol is:
= The sum of index corresponding to number “1” in each column
*Y = The value of index in each
i
experiment: i = 1,…,9.
I
II = The sum of index corresponding to number “2” in each column
III = The sum of index corresponding to number “3” in each column
where
2
2
2
2
9
2
2
2
I
+ II + III
1
9
2
P
3
Q
3
G
--------------------------------
--
--- ------
--
P = I + II + III
Q =
=
G =
Y
S =
–
K =
G
∑
i
3
9
i=1
corresponds to level 2, A3 corresponds to level 3, and similarly for the other factors B, C, and D.
For example, for factor A at levels 1, 2 and 3, the values are
∆tsk(A1) = (3.89 + 2.92 + 2.51)/3 = 3.11°C
∆tsk(A2) = (2.80 + 3.60 + 2.00)/3 = 2.80°C
∆tsk(A3) = (4.38 + 2.09 + 2.83)/3 = 3.10°C
Figure 2 shows that ∆tsk is a function of both different levels and different factors. It is seen
that ∆tsk decreased when either the distance from the outlets or the ambient temperature was
increased, and that it increased when the velocity was increased. The value of ∆tsk is nearly the
same for the three outlet types. The curve for factor C tends to intersect those of factors B and D,
which implies that there may be interactions between B and C and between C and D.
An analysis of variance was conducted using the existing analysis tool of mathematical statis-
tics applied to orthogonal experiments. The results are given in Table 2. In this table the first
five columns contain the same contents as Table 1. The last column lists the resulting Yi (∆tsk)
for each experiment.
As shown in Table 2, the variance value of each factor is SA = 0.18, SB = 2.4, SC = 1.15, and
SD = 1.46. The variance value of interaction is SB×C = SB + SC = 1.64, SB×D = 1.33, and SC×D
=

VOLUME 8, NUMBER 2, APRIL 2002
197
SC + SD = 2.58. Thus, the order of importance of the four factors is B, D, C, and A. The signifi-
cance test was used to judge the significance of the influence of each factor and the interactions.
For factor A (in column 1),
I = 3.89 + 2.92 + 2.51 = 9.32
II = 2.80 + 3.60 + 2.00 = 8.40
III = 4.38 + 2.09 + 2.83 = 9.30
For factor B (in column 2),
I = 3.89 + 2.80 + 4.38 = 11.07
II = 2.92 + 3.60 + 2.09 = 8.61
III = 2.51 + 2.00 + 2.83 = 7.34
For factor A,
P = 9.322 + 8.402 + 9.302 = 243.91
243.91
---------------
Q =
= 81.30
3
For factor B,
P = 11.072 + 8.612 + 7.342 = 250.55
250.55
---------------
Q =
G =
= 83.52
3
9
Y = 3.89 + 2.92 + 2.51 + 2.80 + 3.60 + 2.00 + 4.38 + 2.09 + 2.83 = 27.02
∑
i
i=1
For factor A,
SA = 81.30/3 – 27.022/9 = 0.18
For factor B,
SB = 83.52/3 – 27.022/9 = 2.41
There were no blank columns in the orthogonal table and no repeat values for each test, which
made performing an error analysis difficult. After the variance was calculated, the value of fac-
tor A was found to be so small that it can be neglected. Thus, this factor has little influence on
∆tsk, and SA can be taken as the variance of the error. Furthermore, any interaction relating to
factor A need no longer be considered. The uncertainty of ∆tsk is thus less than 5%.
The results of the calculation of significance test are given in Table 3. The analysis of vari-
ance and significance test show that the order of importance of the factors is B, D, C, and A. Fac-
tor B has a significant influence on the experimental index, factors C and D have some
influence, and factor A has little influence. In addition, interactions exist between factors B and
C and factors C and D.
The analysis shows that the index changes regularly with factors B, C, and D. Because these
three factors are evenly spaced in the experiments, a multiple-unit regression analysis was per-
formed to determine the functional relation between the factor and the index. As shown in Fig-
ure 2, factors B, C, and D are essentially linearly related to the index. A linear regression
equation of the following form was postulated:
ˆ
y = a + bx1 + cx2 + dx3
(3)

198
HVAC&R RESEARCH
Table 3. Significance Test Table
Variance
Source
Partial Sum of
Squares
Degrees of
Freedom
Mean Partial
Sum of Squares
E
Value
Significance
B
C
D
2.40
1.15
1.46
1.64
1.33
2.58
0.18
2
2
2
4
4
4
2
1.2
13.3
6.4
8.1
4.6
3.7
7.2
Significant influence
Slight influence
Slight influence
Slight influence
No influence
0.58
0.73
0.41
0.33
0.65
0.09
B × C
B × D
C × D
Error e(A)
Slight influence
Table 4. Draft Sensation Decomposition Table
Factor Level
B
C
D
1
2
3
5
0
0
0
0
2
5
2
0
where a, b, c, and d are the unknown coefficients and x1, x2, and x3 represent the value of factors
B, C, and D, respectively.
From the regression calculation the values of coefficients were found to be
a = 8.42
b = –2.07
c = 0.83
d = –0.24.
which gives the following relation:
ˆ
y = 8.42 – 2.07x1 + 0.83x2 – 0.24x3
(4)
The coefficient of correlation is 0.957. The regression analysis shows that B, C, and D have a
linear effect on the index ∆tsk.
In order to determine the recommended values for the four factors, the thermal sensation
records were combined with the prior analysis. First, the records were decomposed according to
the importance of each factor. The records of draft sensation for each level of each factor are
given in Table 4.
The draft record at the knees and feet is used. In terms of the total record, the predicted per-
centage dissatisfied PPD for the level with the most uncomfortable environment was calculated.
The values are as follows:
• B1 with a distance of 0.4 m: PPD = 13.9% and the index ∆tsk equals 3.50°C
• C3 with a velocity of 1.5 m/s: PPD = 5.5% and the index ∆tsk equals 3.44°C
• D1 with a temperature of 18°C: PPD = 13.9% and the index ∆tsk equals 3.69°C
When the index ∆tsk is about 3.5°C, the PPD is nearly 14%; when it is about 3°C, the PPD is
approximately 10%. To meet the requirement of an acceptable PPD of less than or equal to
10%, the values in Japan were adopted and a standard of index ∆tsk ≥ °C was set for the evalua-
tion of thermal comfort.
Based on this standard and the regression equation, the recommended values for the factors
were established using Equation 4:
3 ≥ 8.42 – 2.07x1 + 0.83x2 – 0.24x3

VOLUME 8, NUMBER 2, APRIL 2002
199
Thus, the factors must meet the following criteria:
– 5.42 – 0.83x2 + 0.24x3
-----------------------------------------------------------
B = x1 ≥
C = x2 ≤
D = x3 ≥
(5)
(6)
(7)
–2.07
– 5.42 + 2.07x1 + 0.24x3
-----------------------------------------------------------
0.83
– 5.42 + 2.07x1 – 0.83x2
-----------------------------------------------------------
–0.24
Equation 5 shows that if factor C has a maximum value of 1.5 m/s and factor D has a mini-
mum value of 18°C (which are within the range of the experiments), then to provide thermal
comfort, factor B should not be less than 1.13 m. When C equals 0.5 m/s and D = 22°C, then
B has a minimum value of 0.27 m. Using Equation 6 in the same way, the maximum values for
C of 0.33 m/s when B = 0.4 m and D = 18°C, and of 2.3 m/s when B = 1.0 m and D = 22°C, were
established. Equation 7 gives the minimum values for D to be 24.3°C when B = 0.4 m and C =
1.5 m/s, and 15.7°C when B = 1.0 m, and C = 0.5 m/s.
These data were obtained under conditions such that the other factors took on extreme val-
ues. Because such cases would not happen in practical situations, they were readjusted. The
experiments show that when the distance from the outlet is 0.4 m, the value of ∆tsk is 3.69°C,
so the minimum value for this factor was readjusted to 0.5 m. The velocity of supply air nor-
mally cannot be greater than 1.5 m/s, with a maximum value of 2 m/s. The temperature of sup-
ply air should not be below 18°C in view of the vertical temperature gradient. In light of the
interaction between B and C and between C and D, the most extreme values should not be
taken at the same time.
In summary, the recommended values are: the distance between occupant and the outlet is
B ≥ 0.5 m, the velocity of supply air is C ≤ 2 m/s, and the temperature of supply air is D ≥ 18°C.
CONCLUSIONS
The experimental data and analysis of thermal comfort in upward-airflow systems showed
that the four factors that affect thermal comfort in order of importance are (1) distance
between occupant and outlet (factor B), (2) temperature of supply air (D), (3) velocity of sup-
ply air (C), and (4) type of outlet (A). The analysis of variance showed that the distance
between occupant and outlet (B) has a significant influence on thermal comfort, whereas the
type of outlet (A) has little influence on thermal comfort. Interaction exist between the dis-
tance factor (B) and the velocity factor (C) and between the velocity factor (C) and the tem-
perature factor (D).
Based on a multiple linear regression, a functional relation for the skin temperature difference
∆tsk as a function of the distance factor B, the velocity factor C, and the temperature factor D,
was established as given by Equation (4).
The skin temperature difference ∆tsk is the index to use in evaluating the thermal comfort of
an environment. Based on the synthesis analysis and the regression equation, the environment is
comfortable if ∆tsk is less than or equal to 3°C.
The recommended values of three factors (distance, velocity, and temperature) in upward air-
flow were determined. The distance between occupant and the out let (B) should be greater than
or equal to 0.5 m, the velocity of supply air (C) should be less than or equal to 2 m/s, and the
temperature of supply air (D) should be greater than or equal to 18°C.

200
HVAC&R RESEARCH
ACKNOWLEDGMENTS
The project was financially supported by National Natural Science Foundation of China. The authors
also want to express thanks to Ms. A.L. Lian and Ms. Y. Zhang for their translation of the paper.
NOMENCLATURE
a, b, c, d coefficients of regression equation
tforehead forehead temperature
thand
tknee
tsk
hand temperature
knee temperature
skin temperature
variance
A
type of outlet
B
C
D
ET
PPD
tchest
tfoot
distance between occupant and outlet
velocity of the supply air
temperature of the supply air
ventilation efficiency
predicted percentage dissatisfied
chest temperature
S
x1, x2, x3 factors in regression equation
ˆ
y
represents skin temperature difference in
regression equation
foot temperature
∆tsk
skin temperature difference
REFERENCES
Fan, C.Y. 1993. Evaluation Index (PMV) of Thermal Comfort and Its Application. (In Chinese), Journal of
HVAC 23(3).
Fan, C.Y. 1997. The Application of Underfloor Air Conditioning in Office Buildings. (In Chinese), Jour-
nal of HVAC 27(4):30-39.
Fanger, P.O. 1967. Calculation of Thermal Comfort. ASHRAE Transactions 73.
Hanzawa, H. and Y. Nagasawa. 1990. Thermal Comfort of Underfloor Air Supply System. ASHRAE
Transactions 96.
Int-Hout, D. 1990. Thermal Comfort Calculation—A Computer Model. ASHRAE Transactions 96.
Lian, Z.W. and H.Y. Feng. 2002. A Fuzzy Evaluation Model on Indoor Thermal Environment Inside a
Building. (In Chinese), Journal of Shanghai Jiao Tong University 34.
Lian, Z.W. and R.M. Ma. 1996. Effect of the Factors of a Heat Source on HDC in an Upward Displacement
Room. International Journal of Refrigeration 9(3):168-171.
Ma, R.M. 1995. Study on Technology Condition and Comfortable Condition of Underfloor Air Supply
System. (In Chinese), Journal of HVAC 25(6).
Ma, R.M. 1997. The Ventilation Efficiency of Displacement Ventilation and the Evaluation of Its Micro-
thermal Environment. (In Chinese), Journal of HVAC 27(4).
Sodec, F. and R. Craig. 1990. The Underfloor Air Supply System—The European Experience. ASHRAE
Transactions 96:690-695.
Wang, R.X. 1986. Mathematical Statistics. (In Chinese), Xi’an: Xi’an Jiaotong University.