International Journal of Refrigeration 28 (2005) 276287

www.elsevier.com/locate/ijrefrig

Comparison of performances of displacement and mixing

ventilations. Part I: thermal comfort
Zhang Lin*, T.T. Chow, K.F. Fong, Qiuwang Wang, Ying Li
Division of Building Science and Technology, City University of Hong Kong, Hong Kong SAR, Peoples Republic of China
Received 27 May 2003; received in revised form 3 March 2004; accepted 6 April 2004

Abstract
A numerical simulation using a validated Computational Fluid Dynamics (CFD) model of several types of buildings was
used to investigate the difference between mixing and displacement ventilation (DV). The comparison of performances of
displacement and mixing ventilations (MV) will be reported in two parts, one on thermal comfort and the other one on indoor air
quality. This paper, i.e., Part I, compares the performances of floor-supply DV systems with traditional MV systems for offices,
classrooms, retail shops and industrial workshops under a wide range of Hong Kong thermal and flow boundary conditions,
such as a very high cooling load. Through proper design, DV can maintain a thermally comfortable environment that has a low
air velocity, a small temperature difference between the head and ankle level, and a low percentage of dissatisfied people.
q 2004 Elsevier Ltd and IIR. All rights reserved.
Keywords: Air conditioning; Residential building; Ventilation; Simulation; Calculation; Performance; Comfort

Ventilation par deplacement et par melange: comparaison.

Partie I: confort thermique
Mots-cles : Conditionnement dair ; Immeuble dhabitation ; Ventilation ; Simulation ; Calcul ; Performance ; Confort

1. Introduction improvements in the ventilation efficiency at breathing level

are commonly achieved. Areas of stagnant air may be
Indoor spaces, in Hong Kong in particular, are impacted eliminated in an office with partitioned furniture [1],
by a large influx of heat-generating devices characterized by whereas supply diffuser locations at the ceiling may lead
high-density and unevenly distributed loads as well as by the to poor air circulation at the desk in partitioned areas [2].
need for air-cooling throughout the year. A supply air temperature of at least 17 8C is needed to
A system that supplies air at floor level and returns air avoid draught for any method of floor level supply. Room air
near the ceiling has several advantages over conventional is often recirculated into the underfloor plenum to attain this
ceiling supply/return systems. Floor-supply systems dis- temperature level [3,4]. Apart from the vicinity of the supply
tribute conditioned air directly to the occupied zone that is diffusers, the horizontal air temperature variations within
substantially closer to the occupants. With this arrangement, the room seem to be quite small [5]. This has been noted also
when the air temperature in the plenum under the floor
* Corresponding author. Tel.: 852-2788-9805; fax: 852- varied about 3 8C spatially [6]. In order to avoid draught
2788-9716. from the jets, the occupants should be seated at least
E-mail address: bsjzl@cityu.edu.hk (Z. Lin). 1 1.5 m from the supply grilles [4]. No occupant should be
0140-7007/$ - see front matter q 2004 Elsevier Ltd and IIR. All rights reserved.
doi:10.1016/j.ijrefrig.2004.04.005
 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287 277

Nomenclature
A floor surface area (m2)
ce contaminant concentration at the exhaust air (ppm)
cs contaminant concentration at the supply air (ppm)
c contaminant concentration in the room air (ppm)
fcl cloth factor
hc convective heat transfer coefficient between the cloth and air (W/m2 K)
Icl clothing insulation (8C m2/W)
K the turbulent kinetic energy
M metabolism (W/m2)
Pa partial water vapor pressure (Pa)
Qt total cooling load in the room (kW)
Qoe heat generated by occupant, desk lamps, and equipment (kW)
Ql heat generated by lighting (kW)
Qex heat from exterior walls and windows and transmitted solar radiation (kW)
T local air temperature (8C)
Tcl cloth temperature (8C)
Tr mean radiant temperature (8C)
Tu turbulent intensity
u air velocity (m/s)
W external work (W/m2)
h ventilation efficiency
r density (kg/m3)
t mean age of air (s)

located within a radius of the diffuser where the air Designers in Hong Kong are not very familiar with the
velocities are in excess of 0.25 m/s and the temperatures performance of DV.
are more than 0.6 8C lower than the room temperature [7]. A lot of research on DV has been conducted in
The vertical temperature gradient appears to stay within Scandinavian countries [18]. A comprehensive investigation
comfort limits when using a whole-floor supply system [8]. into the DV in the United States was carried by a research
A characteristic feature of the floor air supply system was group [18 21]. The guidelines for designing in the United
the vertical S-shaped temperature pattern [9]. States were also developed [22]. However, direct appli-
Supplying cool air directly to the occupied zone of a cation of the aforementioned results for Hong Kong design
building increases the likelihood of local drafts and may not be feasible because the cooling loads of Hong Kong
temperature stratification. Both of these conditions can buildings are higher than those of US buildings, and much
lead to local thermal discomfort despite the provision of higher than those of Scandinavian buildings. Congested land
acceptable whole-body comfort, as specified by existing use also results in building layouts different from those in
comfort standards [10,11]. An experimental study has the other places. This paper evaluates the performance of
investigated this effect and found some limitations with DV systems on the aspect of thermal comfort under Hong
displacement ventilation (DV) in offices [12]. Kong conditions.
Occupant surveys have found that a thermally comfor-
table environment is among the most important attributes of
a space, but also that this comfort has not been well provided
[13]. Thermal comfort deserves serious inquiry since worker 2. Validated CFD model for research into application of
productivity may be related to satisfaction [14]. In 2001, DV
60% of the electrical energy used in Hong Kong is used in
commercial buildings [15], and a large fraction of this is A computational model was developed based on a
used to provide comfort, thus the efficiency with which commercial CFD program [23]. The model was tested by
comfort is provided is quite important. Lin et al. using the extensive experimental data reported by
In 1989 in Nordic countries, it was estimated that DV Yuan et al. [20].
accounted for a 50% market share of industrial applications Validation was conducted by comparing the flow
and 25% of office applications [16]. The application of floor- patterns, vertical profiles of temperature, concentration,
supply DV system in Hong Kong started in 1980s [17]. velocity, and turbulence intensity between measurement and
However, very few applications can be found in Hong Kong. computation for an individual office, a cubicle office, and a
278 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287

quarter of a classroom, respectively. The computed air 3.1. Evaluation criteria

temperature and velocity agree well with the measured
data [24]. Normal CFD programs do not calculate PD and PPD.
The following models have been implemented in the CFD
program to calculate these parameters [11,26,30]:
PD 34 2 Tu 2 0:050:62 3:14 0:37uTu %
1
3. Performance comparison of DV and mixing
ventilation (for u , 0:05 m=s; use u 0:05 m=s; for PD . 100%, use
PD 100%).

An important objective of air distribution is to create a Tu 1002k0:5 =u%
2

thermal environment with the proper combination of The formulae for calculating PPD can also be found in ISO
comfort variables [25]. The comfort variables are metabolic 7730:1994(E) [11]:
rate, clothing, air velocity, air temperature, air temperature
stratification, radiant temperature, radiant temperature PPD 100 2 95 exp20:03353PMV4 2 0:2179PMV2
asymmetry, relative humidity, and turbulence intensity in
 %
3
the occupied zone [26,27]. For the same level of activity,
clothing type, room geometry, location, and orientation, The PMV in the equation is determined by
thermal comfort is related to air velocity and temperature, PMV 0:303 exp20:036M 0:028
L 4
temperature stratification, relative humidity, and turbulence
intensity. Hence, it is necessary to understand quantitatively with
the local thermal comfort under DV and MV. The aim of L M 2 W 2 {3:96 1028 fcl Tcl 2734
this paper is therefore to analyze and compare the
performances of the two different modes of ventilation 2 Tr 2734
fcl hc Tcl 2 T
mechanism for typical Hong Kong offices, classrooms, retail
shops, and industrial workshops, in terms of thermal 3:05 1023 5733 2 6:99M 2 W 2 Pa

comfort by numerical simulations.
In the ASHRAE comfort standard, air velocity is an 0:42M 2 W 2 58:15 1:7 1025 M5867 2 Pa
important factor in the calculation of the comfort zone. 0:0014M34 2 T} 5
However, the standard also restricts the extent to which air
velocity can be used to achieve comfort, by limiting it to a where the fcl ; Tcl ; and hc are determined by the following
specified maximum of 0.8 m/s in summer [10,28]. Most equations:
researches into the velocity vs. temperature tradeoff have fcl 1:05 0:645Icl for Icl $ 0:078 6
assumed that the temperatures of the ambient air and the air
fcl 1:00 1:290Icl for Icl , 0:078 7
blowing over the skin are the same. One researcher surveyed
a building in which floor-supply ventilation was a source of Tcl 35:7 2 0:028M 2 W 2 Icl {3:96 1028 fcl Tcl
draft, causing discomfort [29]. The transition from desirable
to annoying air movement needs to be studied further, 2734 2 Tr 2734
fcl hc Tcl 2 T 8
particularly for cases where the temperature of the localized The convective heat transfer coefficient, hc ; is determined
air differs from the ambient surroundings. from:
The thermal comfort level is one of the most important
aspects in the performance of a HVAC system. The most hc 2:38Tcl 2 T0:25 for 2:38Tcl 2 T0:25 $ 12:1u0:5 9
common and probably best-understood quantitative method hc 12:1u0:5 for 2:38Tcl 2 T0:25 , 12:1u0:5 10
is predicted mean vote (PMV) for thermal comfort and the
In addition, a CFD program will calculate the distributions
associated percent persons dissatisfied (PPD) [11]. How-
of air velocity and temperature. The thermal comfort
ever, the turbulence of airflow has a significant impact on
provided by DV and MV will be compared in terms of
the sensation of draft, and a mathematical model of draft risk
including turbulence intensity was developed [26]. The
airflow pattern,
model predicts the percentage of people dissatisfied due to temperature distribution,
draft as a function of mean air velocity, turbulence intensity, PD,
and air temperature. The air temperature distribution, the PPD.
percentage of dissatisfied people due to draft (PD), and the
predicted percentage of dissatisfied for thermal comfort The present investigation studies DV systems without
(PPD) are widely used as criteria to evaluate thermal chilled ceiling panels for four types of indoor functional
comfort. spaces: offices, classrooms, retails shops and industrial
 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287 279

Fig. 1. Typical rooms studied: (a) an office, (b) a classroom, (c) a retail shop, (d) a workshop. Fig. 1. Les locaux typiques edudies (a) un bureau,
(b) une salle de classe, (c) un magasin au retail; (d) un atelier.

workshops. Using the same case studies MV was applied larger office has eight sedentary occupants, each with a table
and the results were compared with the DV results. and a computer. A photocopier and printer are also located
Fig. 1 shows the typical layout configurations for each close to the wall. In Fig. 1(b) a typical school classroom is
case study. Fig. 1(a) shows the office layout, which consists shown. There are a total of 40 sedentary occupants
of two smaller individual offices on the left side of the room. (students) and one standing occupant (teacher). A projector
Each of the smaller individual office has on seated occupant is located next to the teacher. A retail workshop is shown in
and a table. A computer is also present on each table. The Fig. 1(c). There are a total of 27 standing occupants. At the

Fig. 2. Supply and exhaust location for DV. Fig. 2. Endroit dapprovisionnement et dechappement pour DV.
280 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287

Fig. 3. Supply and exhaust location for MV. Fig. 3. Endroit dapprovisionnement e dechappement pour la ventilation de melange.

center of the shop floor is a display cabinet. A small office is obtain more meaningful and applicable results a wide range of
also located within the left hand bottom side of the shop boundary conditions were investigated for both ventilation
floor. A PC, Hi-Fi, fax machine and photocopier are also cases. A total of 29 case studies, including 15 office cases, 6
located in this room. Shelves are also present on the walls of cases of classrooms, 6 cases of retail shops and 2 cases of
the retails shop. The workshop is shown in Fig. 1(d), where industrial workshops were investigated. The thermal and flow
eight occupants are standing next to several worktables. boundary conditions are based on typical design conditions in
Additionally an external wall and window is located on the Hong Kong, and are as follows for both mixing and DV:
west wall of all the different room types. The locations of the
supply and exhausts for all four case studies, for mixing and 1.5 ACH # ventilation rate # 31.3 ACH
DV are shown in Figs. 2 and 3. In the MV shown in Fig. 3, 120 W=m2 # Qt =A # 200 W=m2 (550 W/m2 in the case of
the square shapes indicate the supply inlets, while the industrial workshop)
rectangular shapes denote the exhausts outlets. 0:1 # Qoe =Qt # 0:4
The study focuses on the cooling operation because of the 0:3 # Ql =Qt # 0:75
sub-tropical climatic condition in Hong Kong. In order to 0 # Qex =Qt # 0:9

Table 1
Major parameters used in the simulations. Tableau 1. Les parame`ters principaux utilises dans les simulations

No. of West wall Window TV PC Hi-Fi Lamp Photo-copier Machine Air circulation Fresh air intake
person [W] [W] [W] [W] [W] [W] [W] [W] [m3/s] [l/person/s]

Office 10 900 3168 70 10 1280 800 200 0.809 10

Class-room 41 1000 3512 350 1170 0.732 10
Retail shop 27 1096 4000 200 100 200 4700 800 200 1.592 10
Work-shop 8 900 3168 810 33600 0.808 (MV), 10
5.970 (DV)
 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287 281

3.2. Comparison of performances of DV with mixing

ventilation

For the 29 cases, the airflow patterns and the distri-

butions of air temperature, PD and PPD were calculated.
Scenarios of using both a floor-supply DV system and a
conventional MV system for the four types of indoor space
were investigated. The results for the two alternative designs
were compared for each space. The thermal and environ-
mental conditions, including airflow rate and heat source
distribution of each space were identical for both alternative
designs and are listed in Table 1, except that air circulation
for DV is much more in the workshop due to both excessive
cooling load and restriction on temperature gradient. The
DV design uses diffusers installed on the floor and exhaust
grilles on the ceiling, whereas the one MV system has
ceiling mounted diffusers and exhausts grilles for all the four
spaces.

3.3. Airflow pattern

The office is installed with linear diffusers for the two

individual rooms and two rectangular diffusers for the open
office for the DV design. The diffusers are sized so the exit Fig. 4. Airflow pattern for DV. Fig. 4. Mode`le de flux dair pour la
velocity is low. Four rectangular diffusers are used for MV, ventilation de deplacement.
one each for the individual room and two for the large office.
Figs. 4(a) and 5(a) show the airflow patterns in the space for
DV and MV, respectively.
For DV (Fig. 4(a)), it appears that the relatively cold air
from the diffusers is first spreading horizontally upon the
floor, forming a layer of cool carpet, and then slowly rising
to the ceiling exhaust when heated by various heat sources.
The flow near the exterior window and wall rises causing by
the higher wall temperature. The heated objects, such as the
photocopier, the computers and the occupants, generate
strong plumes that bring the air from the lower zone to
the upper zone. The interior partitions are assumed to be
thermally symmetrical and of the same temperature as the
adjacent room air, because all the adjoining rooms are under
very similar indoor climates. Therefore, significant vertical
flows near the interior partitions are not found, as shown in
Fig. 4(a). The air velocity in the room with DV is low
(generally less than 0.2 m/s) except in the thermal plumes
and the flow near the exterior window and wall and the
exhausts at the ceiling.
For MV, the relatively cold air from the ceiling mount
diffusers first falls down to the floor driving by its initial
momentum as well as density difference. The falling flows
behave like jets, which form zones of relative high velocity
in the room. After the jets meet the floor, they first spread
horizontally in all directions, then rises in a similar manner
to that of the DV. The flow near the exterior window and
wall rises in the same manner as that of DV. Thermal
plumes are also generated by the occupants and the office
equipment, which form the upward air flows from the Fig. 5. Airflow pattern for MV. Fig. 5. Mode`le de flux dair pour la
lower zone to the upper zone. The interior partitions are ventilation de melange.
282 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287

treated in the same way as that for DV. The airflow

pattern in the space can be characterized as counter-
current with the supply flows falling down while the
others rising. A typical horizontal section in the occupied
zone has to accommodate both downward and upward
flows, whereas in the case of DV, only upward flows are
found in the occupied zone. The headroom of the office is
limited by the high land cost in Hong Kong. To avoid
draft effect, supply air temperature should not be too
much lower than the room temperature, even for MV.
Hence, the air circulation rates were set identical for both
of the testing cases. The air velocity in the room with MV
might be twice as that of the room with DV if short-
circuiting and entrainment offset each other. The test
result shows that the velocity is generally higher than that
of DV (generally less than 0.3 m/s as shown in Fig. 4(a)).
For DV, the classroom uses one linear slot diffuser
along the exterior wall and eight circular diffusers evenly
distributed on the floor; four ceiling mount rectangular
diffusers are used for MV in the classroom. The airflow
patterns for both DV and MV of the classroom are
similar to those of the office, yet the air velocity for DV
is significantly lower. This is because the cooling load
for the room is lower and leads to smaller air circulation.
The air velocity is generally below 0.1 m/s for DV and Fig. 7. Temperature distribution for MV (8C). Fig. 7. Distribution de
0.3 m/s for MV. Figs. 4(b) and 5(b) show the results for temperature pour ventilation de deplacement (8C).
DV and MV, respectively.

Fig. 6. Temperature distribution for DV (8C). Fig. 6. Distribution de Fig. 8. PD index for DV (%). Fig. 8. Inice PD pour ventilation de
temperature pour ventilation de deplacement (8C). deplacement (%).
 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287 283

Fig. 9. PD index for MV (%). Fig. 9. Inice PD pour ventilation de

melange (%). Fig. 10. PPD index for DV (%). Fig. 10. Inice PPD pour ventilation
de deplacement (%).

Linear slot diffusers are used in the retail shop with DV.
The retail shop with MV also has four rectangular diffusers
3.4. Temperature distribution
mounted on the ceiling supplying air downward. The
generally flow pattern is similar to that of the previous
two types of indoor space. The air velocity is generally Fig. 6(a) (c) show the air temperature for the office, the
lower than 0.2 m/s for DV and 0.3 m/s for MV, respectively. classroom and the retail shop with DV. The air temperature
Thermal plumes are stronger than those found in the office is nearly uniform in the horizontal direction except in the
and the classroom. This is because the internal heat regions close to a heat source in the displacement-ventilated
generation, especially the lighting load, is higher in the office. The supply air is heated first by the floor and mixed
shop. The results are shown in Figs. 4(c) and 5(c). with room air. As a result, the air temperature near the floor
The industrial workshop is designed for the electronic is higher than that of supply air. The phenomenon was also
production, which is installed with equipment of high heat observed by Yuan et al. [20]. To avoid draft, the supply air
generation rate. The cooling load density for the workshop is temperature cannot be too low. The supply air temperature
very high at about 800 W/m2. The correspondent air depends on the cooling load, room geometry and heat source
circulation rate is therefore also high. This results in the type. Generally, the supply air temperature should be in the
highest air velocities observed amongst the four types of range of 20 22 8C for Hong Kong commercial buildings
space, yet the general flow pattern is still similar to those of (Qt =A 120 200 W=m2 ; Ta 24 8C). There is a vertical
the others. The air velocity is generally lower than 0.4 m/s gradient of the air temperature, which depends on the
for DV and 0.6 m/s for MV (Figs. 4(d) and 5(d)). distribution of the heat sources, especially along the height.
The air velocity of DV is generally lower than that of The gradient in the lower part is larger than that in the upper
MV in the occupied zones for all four types of indoor space. part for the office and the classroom where most of the heat
For DV, the velocity in the vicinity of a floor diffuser sources (sedentary occupants) are in the lower part of the
depends on the airflow rate and the face area of the diffuser. rooms. In contrast, larger gradient exists in the upper part of
The layouts of the floor diffusers for the aforementioned the retail shop due to heavy lighting and standing occupants.
four test cases are all different, yet all result in very similar Although the figures show a high air temperature near the
airflow patterns. ceiling, the ceiling surface temperature is several degrees
284 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287

Fig. 11. PPD index for MV (%). Fig. 11. Inice PPD pour ventilation Fig. 13. PPD for MV, office (%). Fig. 13. Inice PPD pour ventilation
de melange (%). de melange (%).

lower. In general, radiation from the ceiling might not be felt

unless the air temperature near the ceiling is very high.
Fig. 7(a) (c) show the air temperature for the office, the
classroom and the retail shop with MV. The temperature
variation along the horizontal direction is larger than that of
DV, which has not been determined to be significant. Other
than this, the temperature distribution patterns for MV are
similar to their DV counterparts.
For the above six cases, the temperature differences
between the head and foot level of a sedentary occupant are
less than 3 K as stipulated in ASHRAE 1992 [10] and
ISO1984 [11].
The temperature uniformity in the horizontal direction is
not maintained if the airflow rate is high. This is evident in
Figs. 6(d) and 7(d) showing the results in the industrial
workshop with DV and MV, respectively. The high
ventilation rate results from high cooling load generated
by the production equipment.

3.5. Percentage dissatisfied due to draft (PD) and predicted

percentage dissatisfied (PPD)

In the occupied zone, the PD is generally less than 10%

Fig. 12. PPD for DV, office (%). Fig. 12. Inice PPD pour ventilation for the office, the classroom and the retail shop with DV and
de deplacement (%). 20% for the 3 spaces with MV. The PD could be greater than
 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287 285

Fig. 14. PPD for DV, for classroom (%). Fig. 14. Inice PPD pour
ventilation de deplacement (%). Fig. 16. PPD for DV, for retail shop (%). Fig. 16. Inice PPD pour
ventilation de deplacement (%).

Fig. 15. PPD for MV, for classroom (%). Fig. 15. Inice PPD pour Fig. 17. PPD for MV, for retail shop (%). Fig. 17. Inice PPD pour
ventilation de melange (%). ventilation de melange (%).
286 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287

Fig. 18. PPD for DV, for workshop (%). Fig. 18. Inice PPD pour Fig. 19. PPD for MV, for workshop (%). Fig. 19. Inice PPD pour
ventilation de deplacement (%). ventilation de melange (%).

25% in the industrial workshop for both modes of people due to draft (PD) is 10% and the predicted
ventilation (Figs. 8 and 9). percentage of dissatisfied is less than 20%.
For DV cases, the PPD is generally less than 20% in the The result shows that besides being not energy-efficient
occupied zone for the industrial workshop, 20% for the and being not cost-efficient, a high ventilation rate might
classroom and 20% for the office and the retail shop. For also have adverse effect on thermal comfort. It was also
MV cases, only the office has the PPD generally lower than found that when there are large heat sources in the room as
20%. All the other 3 spaces have significant proportions of in the case of the workshop, the DV system requires higher
region where the PPD exceeds 20% (Figs. 10 and 11). The air recirculation.
PPD is also shown for two different breathing levels at a
height of 0.6 and 1 m, above the floor level (for sedentary
and standing person relatively) for all four different building
types. This is shown in Figs. 12 19. Acknowledgements
Figs. 8 11 show that only in the vicinity of a floor
diffuser or a major heat source (0.3 m) the PD and PPD are The work was supported by Strategic Research Grant
higher. Otherwise DV outperforms MV. Therefore, DV No. 7001434 of the City University of Hong Kong.
presents a comfort level as good as the conventional MV, if
not better, for typical Hong Kong applications.
References

4. Conclusions [1] A. Hedge, A.T. Michael, S.L. Sharon, Reactions of office

workers and facilities managers to underfloor task ventilation
in offices, J Arch Planning Res 10 (3) (1993) 203 218.
Comparison with conventional MV shows that with
[2] B. McCarry, Underfloor air distribution systems: benefits and
proper design, installation, maintenance and operation, when to use the system in building design, ASHRAE Trans
DV can maintain a thermally comfortable environment. 101 (2) (1995) 902 911.
The air velocity is smaller than 0.2 m/s. The temperature [3] R. Shute, Integrated access floor HVAC: lessons learned,
difference between the head and foot level of a sedentary ASHRAE Trans 101 (2) (1995) 877886.
occupant is less than 3 K. The percentage of dissatisfied [4] F.S. Bauman, E.A. Arens, S. Tanabe, H. Zhang, A. Baharlo,
 Z. Lin et al. / International Journal of Refrigeration 28 (2005) 276287 287

Testing and optimizing the performance of a floor-based task conditioning system with modular services units at the new
conditioning system, Energy Buildings 22 (1995) 173186. Hong Kong Bank Headquarters, ASHRAE Trans 91 (1) (1985)
[5] K. Matsunawa, H. Iizuka, S.-I. Tanabe, Development and 387 403.
application of an underfloor air-conditioning system with [18] X. Yuan, Q. Chen, L.R. Glicksman, A critical review of
improved outlets for a smart building in Tokyo, ASHRAE displacement ventilation, ASHRAE Trans 104 (1A) (1998)
Trans 101 (2) (1995) 887901. 78 90.
[6] H. Fukao, M. Oguro, K. Hiwatashi, M. Ichihara, Environment [19] X. Yuan, Q. Chen, L.R. Glicksman, Models for prediction of
evaluation in an office with floor-based, air-conditioning temperature difference and ventilation effectiveness with
system in an office building, Proceedings of ROOMVENT96, displacement ventilation, ASHRAE Trans 105 (1) (1999)
vol. 3, 1996, p. 31522. 353 367.
[7] K.J. Loudermilk, Underfloor air distribution systems for open [20] X. Yuan, Q. Chen, L.R. Glicksman, Y. Hu, X. Yang,
office applications, ASHRAE Trans: Symp 105 (1) (1999) Measurements and computations of room airflow with
605613. displacement ventilation, ASHRAE Trans 105 (1) (1999)
[8] B.W. Olesen, M. Koganei, G.T. Holbrook, J. Seelen, J.E. 340 352.
Woods, Evaluation of a vertical displacement ventilation [21] S. Hu, Q. Chen, L.R. Glicksman, Comparison of energy
system, Building Energy 29 (1994) 303310. consumption between displacement and mixing ventilation
[9] F. Sodec, R. Craig, Underfloor air supply systemthe systems for different US buildings and climates, ASHRAE
European experience, ASHRAE Trans 96 (2) (1990) Trans 105 (2) (1999) 453 464.
690 695. [22] X. Yuan, Q. Chen, L.R. Glicksman, Performance evaluation
[10] ASHRAE, ANSI/ASHRAE Standard 55-2003, Thermal and design guidelines for displacement ventilation, ASHRAE
environmental conditions for human occupancy, American Trans 105 (1) (1999) 298 309.
Society of Heating, Refrigerating and Air-Conditioning [23] AEA Technology, CFX-4.2 environment user manual, vols.
Engineers, Atlanta, 1992, Meddelande M85: 24. 1 4, 1997.
[11] ISO standard 7730, Moderate thermal environments-determi- [24] Z. Lin, T.T. Chow, K.F. Fong, Q. Chen, Validation of CFD
nation of PMV and PPD indices and specification of the model for research into application of displacement ventilation
conditions for thermal comfort. Geneva: International Stan- to Hong Kong buildings, The Proceedings of the Third
dards Organization; 1994. International Symposium on HVAC, Shenzhen, China, vol. 2,
[12] D.P. Wyon, M. Sandberg, Thermal manikin prediction of 1999, p. 602 13.
discomfort due to displacement ventilation, ASHRAE Trans [25] ASHRAE, Ventilation for acceptable indoor air quality, ANSI/
96 (1) (1990) 6775. ASHRAE Standard 62-1999, ASHRAE, Atlanta, USA, 1999.
[13] Harris, Louis, and Associates, Comfort and productivity in the [26] P.O. Fanger, A.K. Melikov, H. Hanzawa, J. Ring, Turbulence
office of the 80s, The steelcase national study of office and draft, ASHRAE J 31 (7) (1989) 1823.
environments, vol. 2, Steelcase, Grand Rapids, MI, 1980. [27] D. Int-Hout, Thermal comfort calculationsa computer
[14] M. Brill, S.T. Margulis, E. Konar, I. Bost, Using office design model, ASHRAE Trans 96 (1) (1990) 840 844.
to increase productivity, Workplace Design & Productivity, [28] P.O. Fanger, N.K. Christensen, Chart for prediction of
Buffalo, NY, 1984. draught, Aust Refrig Air Cond Heat 40 (7) (1986) 46.
[15] CSD (Census and Statistics Department, Hong Kong S.A.R.), [29] A.H. Huber, Performance of a Gaussian model for centerline
Hong Kong in Figures, 2002. concentrations in the wake of buildings, Atmos Environ 22 (6)
[16] A.G.L. Svensson, Nordic experiences of displacement venti- (1988) 10391050.
lation systems, ASHRAE Trans 95 (2) (1989) 10131017. [30] P.O. Fanger, Thermal comfort, Danish Technical Press,
[17] D. Tuddenham, Design consideration for a floor-based air- Copenhagen, 1970.