Content uploaded by Hyunjae Chang
Author content
All content in this area was uploaded by Hyunjae Chang on Oct 21, 2020
Content may be subject to copyright.
Available online at www.sciencedirect.com
Building and Environment 39 (2004) 153 – 164
www.elsevier.com/locate/buildenv
Eects of outdoor air conditions on hybrid air conditioning based
on task/ambient strategy with natural and mechanical ventilation
in oce buildings
Hyunjae Changa;∗, Shinsuke Katob, Tomoyuki Chikamotoc
aDaewoo Engineering & Construction Co., Ltd, 60 Songjuk-dong Jangan-gu, Suwon, Kyungi-do 440-210, South Korea
bInstitute of Industrial Science, University of Tokyo, 6-1, Komaba 4-chome, Meguro-ku, Tokyo 153-8505, Japan
cNikken Sekkei Ltd, 2-1-2 Koraku, Bunkyo-ku, Tokyo 112-0004, Japan
Received 6 January 2003; received in revised form 26 March 2003; accepted 12 July 2003
Abstract
This research aims to clarify the eects and indoor environmental characteristics of natural and mechanical hybrid air-conditioning
systems in oce buildings during intermediate seasons and to obtain design data. Natural and mechanical hybrid air conditioning is
an air-conditioning system that utilizes natural ventilation and mechanical air-conditioning systems to improve the quality of the indoor
thermal and air environment, and to reduce energy consumption. This report rst categorizes the available natural ventilation conditions
and estimates the amount of natural ventilation available in a model building. Furthermore, based on the concept of task-ambient air
conditioning, after controlling the average temperature in the task zone to a target air conditioning temperature (26◦C), changes in the
outdoor temperature/humidity and the inow rate, and the indoor environment and amount of cool heat input were studied with changes in
the size of the natural vent using three-dimensional Computational Fluid Dynamics (CFD) analysis. The results of these studies indicated
that natural ventilation at temperatures lower than the indoor temperature eectively covered the lower indoor task zone through negative
buoyancy, which enabled energy-saving air conditioning in the task zone.
?2003 Elsevier Ltd. All rights reserved.
Keywords: Natural ventilation; CFD; Ventilation eciency
1. Introduction
Outdoor air cooling and natural ventilation through open
windows have both been proposed and actually used as
energy-saving cooling methods for oces. However, the for-
mer is primarily used to remove the heat load and introduce
outdoor air when the heating is turned o during interme-
diate seasons. Direct introduction of outdoor air by opening
the windows has been developed signicantly in past stud-
ies. For an example of recent accomplishments on this, Hunt
and Linden describe the uid mechanics of natural venti-
lation by the combined eects of buoyancy and wind [1].
W.Z. Lu et al. investigated the characteristics of cross ven-
tilation in a designated refuge oor in a high-rise building
[2]. Yuguo Li et al. derived analytical solutions for cal-
culating natural ventilation ow rates and air temperatures
∗Corresponding author. Tel.: +82-31-250-1216;
fax: +82-31-250-1131.
E-mail address: changhj@ihanyang.ac.kr (H. Chang).
in a single-zone building [3,4]. M.M. Eftekhari et al. studied
air ow distribution in and around a single-sided naturally
ventilated room [5]. However, direct introduction of outdoor
air through the windows does not necessarily guarantee con-
trol of the indoor thermal environment within a comfortable
range, because it allows air to enter, but gives no control
over the temperature of the air inow from outdoors or the
ow rate, and the period during which windows can be left
open is often limited. Thus, this research aims to develop a
new natural and mechanical hybrid air-conditioning system
that will maintain a comfortable indoor thermal environment
and maintain air quality, while achieving energy conserva-
tion by a rational use of natural ventilation, achieved by
opening natural indoor vents such as windows, and mechan-
ical cooling. For this purpose, the concept of task-ambient
air conditioning was introduced, which is founded on the
need for an even distribution of indoor temperatures and
air quality while breaking away from the conventional
fully mixed indoor air conditioning. This adopts the con-
cept illustrated in Fig. 1, where the heat and contaminants
0360-1323/$ - see front matter ?2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.buildenv.2003.07.008
154 H. Chang et al. / Building and Environment 39 (2004) 153 – 164
Cool fresh air
Thermal stratification
being formed
Efficient exhausting
of excess heat
Ambient zone
Task one
Exhausting
contaminants
Comfortable in terms of heat andair quality
z
Fig. 1. Concept of natural and mechanical hybrid air conditioning system.
generated in an oce indoor task zone are exhausted to
the outdoors by natural ventilation such as opening win-
dows, while using air conditioning to assist with removing
part of the heat load when this is unachievable by natural
ventilation alone. Compared to ordinary oces, which in-
troduce only a minimum amount of outdoor air as required,
those with a natural and mechanical hybrid air-conditioning
system can expect to have a tremendous increase in the
amount of outdoor air introduced, and this will signicantly
improve the indoor air quality when using natural venti-
lation. This was ascertained from eld measurements in
Liberty Tower at Meiji University, which is located in the
center of the Tokyo Metropolitan area in Japan, and which
has introduced a hybrid air-conditioning (cooling) system
[6]. The natural ventilation system in Liberty Tower is de-
signed to introduce outdoor air through an opening under
the window. The outdoor air introduced rises up through
the wind core (escalator void) by the stack eect, and is
then exhausted at the wind oor on the 18th level. From
the results of the eld measurements, however, in order to
make feasible a ventilation and air-conditioning system that
is most likely to provide these energy-conservation eects
and good air quality, it is critical to determine the mixture
and dispersion properties of natural ventilation air and in-
door air conditioning air, as well as to estimate the amount
of natural ventilation available. At present, there are almost
no estimation or evaluation methods that take these into
consideration. This research carries out a detailed study of
the following in providing oce space with natural venti-
lation by opening indoor vents in the ambient zone while
cooling the task zone:
1. Indoor micro heat/air mixture-dispersion phenomena by
Computational Fluid Dynamics (CFD).
2. The indoor thermal environment and ventilation e-
ciency.
3. Heat transfer phenomena in the task and ambient zones.
4. The energy-conservation eects (indoor load reduction
and reduction in the power of the air-conditioner fan
throughout the year).
A comprehensive examination is conducted into the eec-
tiveness of a natural and mechanical hybrid air-conditioning
system and what constitutes an ecient air-conditioning sys-
tem. This report describes the concept of task-ambient hy-
brid air conditioning, followed by the results of examining
the eects of outdoor air conditions on energy conservation.
2. Method of analysis
The analysis ow for this research is shown in Fig. 2. The
amount of natural ventilation in a single room was estimated
Macro model Micro model
The indoor air current and temperature
distribution and ventilation efficiency for
an office with a hybrid air conditioning
system were analyzed with the amount of
natural ventilation based on the results of
macro model 1. With the average
temperature in the task zone maintained at
a constant 26˚C, the following were
conducted:
- CFD analysis of the indoor air current
and thermal environment in the office
with a hybrid air conditioning system
- Analysis based on CFD results for indoor
ventilation efficiency in the office
From micro model analysis results
- Analysis of mixture rate of natural
ventilation and air conditioning air
- Analysis of the amount of heat mixed
and dispersed between the task zone
and the ambient zone
Evaluation of a hybrid air conditioning system
- Evaluation of indoor heat and air quality distribution properties in
the office
- Evaluation of indoor energy in the office
Macro model 1:
The amount of
natural ventilation
in the entire buil-
ding was comput-
ed by ventilation
circuit network
analysis.
Macro model 2:
The annual ene-
rgy simulation to-
ol based on a
block model was
made by using a
database contain-
ing the amount of
indoor heat load
removed by natu-
ral ventilation and
air conditioning.
Fig. 2. Research ow.
H. Chang et al. / Building and Environment 39 (2004) 153 – 164 155
Y axi s:
Assuming the reference
value of the summer cool-
ing load to be 70(W/m2),
the amount of natural
ventilation required to
remove this load
[m3/(h.m2)]
45 [m3/(h.m2)] (assumed
to be air of a volume not
exceeding a wind velo-
city of 1.5 [m/s] (assum-
ption) which may blow
papers off the desk*))
7.5 [m3/(h.m2)]
(minimum OA: Outdoor
air required per person
30 [m3/(h person)]
x Occupancy 0.25
[person/m2]) The lowest limit at which
the possibility of condensa-
tion is considered
Established temperature (the
upper limit of outdoor temp-
erature effective for remov-
ing the cooling load)
X axis:
Outdoor temperature [˚C]
Equation for volume of natural ventilation air required :
70[W/m2]
Cp.ρ: 0.34 [(W.h)/(˚C.m3)] × (Indoor Temp. 26[˚C] - Outdoor Temp. [˚C, X axis])
Conditions of excessive
natural ventilation
Natural ventilation air
volume being tighten-
ed by controlling
the ventilation
air volume
through the
opening
Portion of load removed by
air conditioning assistance
Conditions of insufficient
natural ventilation
Portion of load removed
by natural ventilation
Upper limit of ventilation
amount (in order to prevent
papers from being blown
away)
Frequency of ventilation
15 [times/h]
10 [times/h]
5 [times/h]
Amount of outdoor ai
r
introduced by ordinary ai
r
conditioning
50
40
30
20
10
014 16 18 20 22 24 26 28
Fig. 3. Concept of control by a natural and mechanical hybrid air conditioning system.
∗Computed as the instantaneous wind velocity on desks 1.5 m/s×wind velocity ratio on window surfaces and desks 1.2 (assumption on safety side)/G.F.2
(assumption)/oce depth 10.8 m×opening vertical width 0.5 m×ow coecient 0.3×3600.
by wind-driven ventilation macro-analysis for a model build-
ing, followed by micro-analysis of the indoor temperatures
and air quality distribution properties to control the amount
of heat transfer in the indoor task and ambient zones, based
on which macro-analyses were conducted on the air condi-
tioning energy required annually.
2.1. Conditions of available natural ventilation
The results of organized conditions for natural ventila-
tion made available by a natural and mechanical hybrid
air-conditioning system are shown in Fig. 3. Available natu-
ral ventilation is determined by the outdoor temperature and
the air volume introduced indoors. Outdoor temperatures are
shown on the horizontal axis; the amount of natural ven-
tilation per unit area [m3=(h m2)] is shown on the vertical
axis. Ventilation rates for the oce model adopted for this
research are shown at the right of the gure. The graph indi-
cates the amount of natural ventilation required to remove an
indoor cooling load—(70 W=m2) in this case. The oce in-
doors is assumed to have fully mixed air conditioning. More
outdoor air becomes necessary when the outdoor tempera-
ture becomes higher, and the cooling load can be adequately
reduced with less outdoor air when the outdoor temperature
becomes lower. When there is too much inow of outdoor
air, problems occur, such as papers being blown o desks,
and when outdoor temperatures are too low (16◦C or lower
for example), condensation and cold drafts occur, making
natural ventilation unusable. The area above the graph shows
excessive (adequate) natural ventilation conditions, and the
area under the graph shows conditions in which the load re-
moved by natural ventilation is insucient. This research
Targeted mid- to high-rise building:
floor height at 4m
25-story building
(100m)
VOID
Office space on each floor
(open to exterior and void sides)
Office depth: void depth:
10.8m 10.8m
2-dimensional analysis
(continuous in depth direction)
Exterior wind velocity: 3.0m/s
vertical ventilation width: 0.5m;
wind direction: 90˚ to window surface
Fig. 4. Oce building model.
was intended to show the use of mechanical cooling in as-
sisting conditions where natural ventilation is insucient.
2.2. Macro model 1
The assumed oce building model has a void (court-
yard) in the center as shown in Fig. 4, and is a mid- to
high-rise oce building with a cooling load in the interior
area throughout the year. The amount of natural ventilation
was analyzed by a ventilation network computation, taking
assumed values for the outdoor wind direction, wind velo-
city, wind pressure coecient, and opening ow coecient.
The amount of indoor natural ventilation targeted for study
was obtained through this analysis and used as the boundary
conditions for the micro model analysis.
156 H. Chang et al. / Building and Environment 39 (2004) 153 – 164
2.3. Micro model
Based on the amount of natural ventilation obtained, a
CFD analysis was conducted on the thermal and air quality
distribution in the oce. Based on these results, indoor
air quality was analyzed using a scale for ventilation
eciency (SVE) such as the spatial distribution of the age
of fresh outdoor air (SVE 3) and the contribution ratio of
the natural ventilation inlet (SVE 4) [7]. The thermal envi-
ronment and air quality of natural and mechanical hybrid
air conditioning based on a task-ambient air condition-
ing concept that diers from the indoor fully mixed air
conditioning method were investigated by these analyses.
CFD analyses were conducted by a steady-state compu-
tation, and time uctuations in natural ventilation were
not taken into consideration. Fluctuation components in
natural ventilation have a signicant inuence on dealing
with human comfort, aspects of which are currently being
studied [8].
2.4. Macro model 2
From the above results, the macro properties of heat and
air that transfer and disperse between the task, ambient, and
perimeter areas were analyzed and reanalyzed by annual
simulation based on a macro model. However, the results of
the macro model 2 annual simulation are not described in
this report; they will be reported subsequently.
3. Establishing the oce model
3.1. Establishing the model
The oce model established in this research is shown
in Fig. 5. It was assumed to be a 10.8 m deep continuous
oce, with a 1:8 m wide section (half of the 3:6 m space)
used for analysis. One side of one wall (the left wall sur-
face in Fig. 4) faced the exterior, with the other wall facing
the void (Fig. 3), both of which had an opening for natu-
ral ventilation above the window. Natural ventilation was
Exhaust
PC.
Pa
Human
model rtition
Wall
Ceiling
Chair
Symmetry
plane
:
Natural Ventilation
x
1
x
2
x
3
Unit
:
m
Window
glass
Lighting (0.2 × 1.2)
&
Exhaust opening (0.1 × 1.2
)
Floor supply opening
(0.2
×
0.1)
Natural Ventilation
Exhaust window opening
2.6
0
.
5
Supply
:
Mechanical Air-conditioning
Desk
Floor
Natural Ventilation
Suply window Opening
Natural ventilation
supply window opening
Natural ventilation
exhaust window opening
Mechanical air-conditioning
Natural ventilation
Fig. 5. Targeted oce (for micro analysis).
induced by negative pressure at the top of the void, allow-
ing outdoor air to ow in from the outside (the left wall
surface in Fig. 4) and pass through to the void. Air for
a hybrid air-conditioning system was supplied through air
diusers in the oor intended to provide task air condi-
tioning. Air was exhausted from the ceiling return. Also,
fresh outdoor air was introduced to the indoors only by nat-
ural ventilation, and was not introduced by air condition-
ing through air diusers. The oce was divided into 3:6m
spaces by I-form partitions perpendicular to the window
sides, and in which desks were placed. Each desk had a
surface generating heat from a PC, etc., as an indoor heat
source.
3.2. Dividing the targeted oce into domains
The oce was divided into three domains: task, ambient,
and perimeter zones, the thermal characteristics for each of
which were analyzed (see Fig. 6). The human habitation
area was the main target of control in this research, and
was named the task zone. With the task zone as a start-
ing point, the perimeter area was established as a space
up to 0:9 m from the glass surface with a solar radiation
load. The interior area was divided into task and ambient
zones; the task zone considered as a space for humans to
work while seated, with a domain extending 1:5 m above the
oor.
3.3. Establishing indoor heat loads
Heat loads in the analysis spaces were established as
shown in Table 1.
1. Lighting load: 400 W; from lighting sources on the
ceiling.
2. Solar radiation load: to be uniform on the inside surface of
the exterior window assuming the amount of south-facing
all-weather solar radiation in Tokyo at 14:00 in May
(200 W=m2) to be 0.5 = transmittance of the glass +
absorptivity (with the assumption that sunlight does not
strike the void side).
H. Chang et al. / Building and Environment 39 (2004) 153 – 164 157
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
EEEEE E
0 10 20 30 30 20 10 0
Exterior wind velocity
in upper air: 3m/s
Windward office
Indoor heat generated: 87W/m2
Boundary conditions for
windward opening
Opening vertical width: 0.5m
Wind pressure coefficient: 0.7
Flow rate coefficient: 0.3
25F
20F
15F
10F
1F
5F
18h-1
23h-1
26h-1
29h-1
32h-1
33h-1
8h-1
12h-1
19h-1
24h-1
29h-1
32h-1
Void top boundary conditions:
Opening size: 10.8m
Wind pressure coefficient: -0.4
Flow rate coefficient: 0.8
Void side opening boundary
conditions:
Vertical opening width: 0.5m
Flow rate coefficient: 0.1
(Obstacles such as hallways
were considered.)
Indoor air
flowing out
Leeward office
Outdoor air
flowing in Leeward opening boundary
conditions:
Vertical opening width: 0.5m
Wind pressure coefficient: -0.5
Flow rate coefficient: 0.3
Ventilation frequency (h-1)Ventilation frequency (h-1)
VOID
Fig. 6. Examples of ventilation circuit network.
Table 1
Cooling loads of domains for computationa
Heat generation Lighting Solar radiation PC Human model Floor surface Total
unit (4 units) (window surface) (4 units) (1 unit) (human)
Amount 400 225 800 55(SH) 220(SHb) 1700(SH)
of heat 28(LH) 111(LHc) 139(LH)
(W)
aSensible heat load per oor area (19:4m
2)is87:6[W=m2].
bSH: sensible heat.
cLH: latent heat.
3. Oce automation equipment load: 4 PC units per desk
(800 W total).
4. Humans: approximately 0.25 persons/m2occupancy
load; in order to evaluate the thermal environment with
respect to heat loss/gain by convection and the radiation
between humans and the surrounding environment, one
human model was placed in the space to be analyzed;
other human heat loads were uniformly positioned on
the oor (total of both sensible heat loads: 275 W); a
0:45 m ×0:33 m ×0:88 m block was used as a human
model for the convenience of the CFD analysis (see
Fig. 4).
In the CFD analysis, the distribution of the indoor cooling
load convection transfer was provided by a coupled simu-
lation of radiation/convection made using preparatory com-
putations. 1
1Boundary conditions on simulated heat ows were computed by giving
the amount of convection transmission heat obtained through preparatory
computations by a radiation and convection coupled analysis. Strictly
speaking, the distribution of the amount of convection transmission heat
diered in each case; however, identical distribution properties were
assumed because dierences in the indoor heat loads were not so great,
and the indoor environmental properties were not thought to vary so
signicantly either.
4. Computing the ventilation rate
Assuming a 25-story mid- to high-rise oce building, the
rate of ventilation from the wind pressure and buoyancy was
analyzed based on the ventilation network. The assumed
wind pressure coecient, ow coecient at the opening,
and sample results are shown in Fig. 6. The ventilation rate
was computed with the outdoor wind velocity assumed to
be 1.0–3:0m=s, the ventilation window vertical width to be
0.1–0:5 m and the wind direction to be varied at 45 –90◦to
the window surface. As a result, a ventilation rate of 15–30
times/h was obtained at mid-level in the oce building on
the leeward as well as windward sides due to negative pres-
sure (wind pressure coecient: −0:4) 2at the void top (ap-
proximately 30 times/h at lower level; 5–20 times/h at upper
level). Some reverse ventilation was obtained at the upper-
most level on the leeward side. Results varied depending
on dierences in the conditions noted above. In establish-
ing boundary conditions for analyzing the properties of the
indoor thermal environment distribution by micro-analysis
(CFD analysis), the use of air conditioning and a uniform
2The wind pressure coecient was assumed to be constant on the
building surface. Also, the outdoor wind velocity distribution was con-
sidered only in the vertical directions (1/4 power law).
158 H. Chang et al. / Building and Environment 39 (2004) 153 – 164
ventilation rate of 10 times/h for natural ventilation were
established, and further micro-analyses were conducted.
5. Micro-analysis
5.1. Method of analysis
(1) Indoor ow-eld simulation. Flow-elds were
simulated by three-dimensional CFD analysis based on
the standard k–model [9]. The continuity equation
and the fundamental equations governing the motion of
steady, incompressible, and turbulent ows are the
averaged Navier-Stocks equations that can be expressed as
follows:
@Ui
@xi
=0;
Ui
@Ui
@xi
=−1
@P
@xi
+@
@xi@Ui
@xj
+@Uj
@xi−u
iu
j:(1)
The k–model relates the turbulent kinetic energy kand the
dissipation rate of turbulence by
vt=C
k2
;
u
iu
j=−vt@Ui
@xj
+@Uj
@xi+2
3kij;(2)
where ij is the Kronecker symbol.
Kand can be solved by the following transport equa-
tions:
Uj
@k
@xj
=@
@xjv+vt
k@k
@xj+Pk−;
Table 2
Boundary and computation conditions
Air diuser •Uin is the velocity at the natural ventilation inow and air conditioning air diuser
•kin =3=2(Uin ×0:05)2
•in =C×k3=2
in =lin
•lin =1=7 of the air diuser width
•Tin :Tsupply
•AHin :AH
supply
Air return Velocity is free outow (based on the law of conservations of mass)
kout,out ,T, AH=Free slip
•Velocity: wall surface is based on generalised logarithmic law; symmetry surfaces are free slip.
Wall surface •Heat: the amount of convection heat transfer is xed
•Absolute temperature,
◦human model, oor surface: the amount of heat generated is constant
◦Other walls: humidity gradient =0
Analysis meshes 88 (X1)×17 (X2)×15 (X3)=22;440
Note:U: Air current velocity (m=s); k: energy of turbulence (m2=s2); : dissipation rate of k(m2=s3). lin : air diuser turbulence scale (m); T:
temperature (◦C); AH: absolute humidity (kg=kg).
Uj
@
@xj
=@
@xjv+vt
@
@xj+C1Pk−
k−C2
2
k;
Pk=vt
@Ui
@xj@Ui
@xj
+@Uj
@xi:(3)
The model constants are as follows:
C=0:09;C
1=1:44;C
2=1:92;
k=1:0;and =1:3:
A CFD simulation based on the standard k–model was
validated by comparing the results of the room airow sim-
ulation with the experimental results [10,11]. Detailed CFD
boundary conditions are shown in Table 2.
(2) Air-conditioning system control. The variable air
volume (VAV) method was assumed for controlling
oor-diused air conditioning in the task zone. The aver-
age temperature in the task zone for each CFD step was
obtained, and the dierence from the air conditioning target
(average task zone temperature: 26◦C) was quantitatively
computed. The air conditioning air volume was gradually
changed according to the dierence, which was computed
again by CFD. The volume of air conditioning air and
indoor input heat from air conditioning when the air con-
ditioning target was achieved were analyzed through these
convergence computations.
5.2. Analysis cases
Analysis cases are shown in Tables 3and 4. In the oce
model where the average space temperature in the task zone
is controlled by air conditioning at the target temperature
(26◦C), the indoor thermal and air quality environment and
the air conditioning input heat volume were analyzed for the
H. Chang et al. / Building and Environment 39 (2004) 153 – 164 159
Table 3
Analysis conditions and results (total of indoor heat generated targeted for computation: 1700 W; oor area = 19:4m
2)
Air conditioning system Natural ventilation Temperature
when
Cooling Imposed Diused Air Temperature/ Amount Imposed Inow Air Temperature/ Air Ar of instantaneous
load load air volume humidity of fresh load wind ow humidity inlet natural and uniform
Case (W)(∗) velocity (∗)(
◦C)/(%) outdoor (∗) velocity rate (◦C)/(%) width ventilation dissipation
(W)(
∗)(m
3=h) air (W) (m/s) (m3=h) (m) air inow is assumed
(m/s) introduced (∗)(
∗)
(m3=h) (◦C)
A 0 0 0 1700 18.8/69 −4:83 28.6
B 146 0.07 25.2 1554 19.5/66 −4:36 28.8
C 366 0.31 111.6 1334 0.16 10 21.0/60 −3:36 28.7
D 627 0.49 176.4 1073 22.5/54 −2:35 28.9
E 843 0.66 237.6 19.0 0 857 24.0/50 0.5 −1:34 29.1
F 1700 882 0.64 230.4 /80 818 0.08 5 −13:42 30.3
G 563 0.39 140.4 1137 0.12 7.5 −5:97 30.0
H 129 0.09 32.4 1571 0.20 12.5 21.0 −2:15 28.4
I 0 0 0 1700 0.23 15 /60 −1:49 27.6
J 491 0.46 165.6 1209 0.39 10 0.2 −0:21 27.9
K 545 0.55 198.0 1155 0.78 10 0.1 −0:03 27.5
Note:(
∗) was obtained from results of CFD analysis.
Table 4
Analysis cases when the temperature of natural ventilation air ow varied
Cases C-1 C-2 C-3 C-4 C-5 C-6 C-7
Inow outdoor relative humidity
(%) 30405060708090
Note: Conditions other than inow outdoor relative humidity in Table 4are the same as Case C in Table 3.
following cases:
1. Inow air temperature varied assuming that the natural
ventilation inow rate (10 times/h) and absolute humidity
were xed (Cases A–E).
2. Air inow rate varied assuming the natural ventilation
temperature remained constant (21◦C) (Cases F–I).
3. Size of the natural ventilation inlet varied assuming that
the natural ventilation temperature and ow rate were
xed (Cases J–K).
4. Inow humidity varied assuming that the inow
temperature (21◦C) and natural ventilation inow
rate (10 times/h) were xed (Table 2, Cases C-1
to C-7). Case C is the reference case when the
air inow rate, size of inlet, and inow humidity
varied.
5.3. Analysis results
(1) When the inow temperature varied (Cases A–E).
Only results for Cases A, C, and E are shown.
(a) Air current distribution. The air current distribution
when the inow temperature varied is shown in Fig. 7. The
ow patterns of the natural ventilation inow jets indicated a
decrease in the Ar number 3with an increase in the outdoor
temperature, and the jets penetrated deeper into the interior.
Because the dierence between the outdoor and indoor tem-
peratures was great in Case A (Ar number = −4:83), the
inow air fell along the walls on the window side due to the
strong negative buoyancy, and owed towards the inside at
a slow speed. Air that reached the wall surface on the other
side rose along the wall surface, and was discharged from
the natural ventilation outlet. Because the target air condi-
tioning temperature (26◦C) was achieved only by natural
ventilation in Case A, the wind velocity at the air condi-
tioning air diusers became zero. The natural ventilation in-
ow jets penetrated deeper into the interior and mixed bet-
ter with the indoor air in Case C (Ar = −3:36) and Case E
(Ar = −1:34) than in Case A.
(b) Temperature distribution. The temperature distribu-
tion when the inow temperature varied is shown in Fig. 8.
Because inow air at a relatively low temperature (18:8◦C)
owed into the interior without mixing very well with the
indoor air in Case A, temperature stratications were formed
3Ar = g L=U 2; where g: gravitational acceleration [m=s2]; : ex-
pansion coecient [◦C−1]; : temperature dierence between the natural
ventilation inlet and the average task zone [◦C]; L: natural ventilation
inlet height [m]; U: natural ventilation inow velocity [m/s].
160 H. Chang et al. / Building and Environment 39 (2004) 153 – 164
(b) Case C (Ar number of inflow outdoor air = -3.36; ∆T = 5.0°C)
Ceiling return
21.0°C
24.0°C
(c) Case E (Ar number of inflow outdoor air = -1.34; ∆T = 2.0°C)
1 m/s
Natural ventilation inlet
0.16m/s, 18.8°C
Perimeter zone
Human model Ambient zone
Tas k zone
(a) Case A (Ar number of inflow outdoor air = -4.83; ∆T = 7.2°C)
Natural venti-
lation outlet
(∆T: (Average task zone temperature (26°C)) - (Outdoor temperature))
Fig. 7. Air current distribution (center sections including human models;
air diusers installed in the oor).
in the general space. Also, the air conditioning ow rate
(temperature at air diuser: 19◦C) increased with an increase
in the natural ventilation inow temperature to maintain the
average temperature (26◦C) in the task zone, and the tem-
perature gradient near the oor became great.
(2) When the inow rate varied (Cases F–I)
(a) Air current distribution. Figures for air current dis-
tribution when the natural ventilation inow rate varied are
omitted because of their similarity with Fig. 6. The ow pat-
terns of the natural ventilation inow jets varied intricately
with the increase in the outdoor air inow rate. Natural ven-
tilation vents were continuous in a slot diuser congura-
tion, and the jets exhibited a two-dimensional aspect when
the natural ventilation ow rate was high; however, they did
not necessarily become two-dimensional when the ow rate
was low, and exhibited a three-dimensional aspect when in-
uenced by air currents from the indoor air conditioning air
diusers. The natural ventilation jets varied in a complex re-
lationship to many factors, such as negative buoyancy acting
on the surface of the ceiling, the walls and jets themselves,
the fact that the air conditioning diuser jets decrease the
air ow rate in inverse proportion to an increase in the nat-
ural ventilation ow rate, and the indoor circulating ow
properties.
(b) Temperature distribution. Figures for the temper-
ature distribution when the natural ventilation inow rate
varied are excluded because of their similarity with Fig. 7.
(a) Case A (a : 26.2°C, b : 26.0°C)
26
26
27
25
27
28
Natural ventilation inlet
0.16m/s, 18.8°C
Perimeter zone Human model Ambient zone
Task zone
Natural venti-
lation outlet
a: Average indoor temperature
b: Average task zone temperature
(b) Case C (a : 26.6°C, b : 26.0°C)
Ceiling return
21.0°C
26
26 25
27
27
26
27 25
24.0°C
(c) Case E (a : 26.9°C, b : 26.0°C)
26
26
26
27
25
24
27
27
28
23
27
Fig. 8. Temperature distribution (center sections including human models;
air diusers installed in the oor).
Relatively stable temperature stratications were formed in
task-ambient zones that the natural ventilation jets did not
reach. The high/low temperature gradient became smaller
with the increase in the natural ventilation air inow rate,
reducing the relative high temperature zone near the ceiling.
This was thought to be because a large quantity of natural
ventilation air, hotter than the current from the air condi-
tioning air diusers, owed into the task zone and mixed
well with the surrounding air.
(3) When the inlet size varied
(a) Air current distribution. Figures for air current dis-
tribution when the air inlet size varied are omitted because
of their similarity with Fig. 6. As the inow velocity of
the natural ventilation air became greater with a decrease in
the height of the natural ventilation inlet (with temperature
and ow rate constant), the natural ventilation inow jets
penetrated deeper into the interior. The natural ventilation
inow jets owed into the interior along the ceiling sur-
face, and fell where the human model was located in Case
J (Opening height: 0:2 m; Ar number at natural ventilation
inlet: −0:21). The natural ventilation inow jets passed the
human model in Case K (Opening height: 0:1 m; Ar num-
ber: −0:03). The velocity of the natural ventilation inow
in Cases J and K exceeded 0:5m=s immediately after the
natural ventilation inlet; however, it reduced to a low speed
of 0:25 m=s or slower near the human model in each case.
H. Chang et al. / Building and Environment 39 (2004) 153 – 164 161
Fig. 9. Relative humidity distribution (center sections including human
models; air diusers installed in the oor).
(b) Temperature distribution. Figures for the tempera-
ture distribution when the air inlet size varied are excluded
because of their similarity with Fig. 7. No outstanding tem-
perature stratication was observed in the general space in
any case. However, as the inow velocity of the natural ven-
tilation air became greater with a decrease in the height of
the natural ventilation inlet and the natural ventilation inow
air penetrated deeper into the interior, mixing well with the
indoor air, the temperature gradient became small near the
oor on the right side of the room across from the air inlet.
Also, the isotherms became parallel and indicated a slight
tendency to form stratications on the left side of the room.
(4) When the inow humidity varied. Only the results for
Cases C-1, C-3, C-5, and C-7 are shown.
(a) Relative humidity distribution. The indoor relative
humidity distribution when the natural ventilation humidity
varied is shown in Fig 9. As the ow rate of the natural ven-
tilation inow air was higher than that at the air conditioning
Fig. 10. PMV distribution (center sections including human models; air
diusers installed in the oor).
air diusers, the indoor relative humidity was inuenced by
the outdoor temperature. Also, the indoor distribution was
signicantly inuenced by the temperature distribution ((b)
in Fig. 7). Cases with the natural ventilation inow at a 50%
or lower relative humidity (Cases C-1, C-2, and C-3), gen-
erally indicated a humidity of 50% or lower. Cases with the
inow at a 70% or higher relative humidity (Cases C-5, C-6,
and C-7), generally exceeded a relative humidity of 50%.
(b) PMV distribution. The PMV distribution 4when the
inow outdoor humidity varied is shown in Fig. 10. The
indoor PMV distribution properties were signicantly inu-
enced by the temperature distribution ((b) in Fig. 7). The
average PMV values in the task zone were 1.4–1.7 in all
analysis cases, which felt warm. Also, the PPD then was
46– 62%. This is thought to have been primarily due to great
4The fundamental heat resistance value of clothing was assumed to be
0.6 (clo), and the amount of metabolism to be 1.2 (met). The wall surface
temperature was obtained from the temperature near the wall surface and
the amount of convection heat transfer, assuming the convection heat
transfer to be 4 (W=m2◦C), from which the average radiation temperature
was computed.
162 H. Chang et al. / Building and Environment 39 (2004) 153 – 164
1
0.8
0.8
1
(a) Case A (τn =360sec)
Natural ventilation inlet
0.16m/s, 18.8°C
Perimeter zone
Human model Ambient zone
Tas k zone
Natural venti-
lation outlet
Ceiling return
(b) Case C (τn =360sec)
21.0°C
1
0.8
0.8
0.6
0.4
0.2
(c) Case E (τn =360sec)
1
1
1.2
1
0.8
0.6
0.6
0.8
0.4
24.0°C
Fig. 11. Non-dimensional age of air distribution (center sections including
human models; air diusers installed in the oor).
interior heat generation and a high average radiation tem-
perature. The PMV became greater with an increase in the
natural ventilation inow humidity; however, while the in-
owing outdoor air relative humidity varied from 30% (Case
A) to 90% (Case G), the average PMV value was more than
0.3, from 1.4–1.7, which was conrmed comparatively in-
sensitive to the relative humidity.
(5) Age of air and contribution ratio
(a) Age of air (SVE3). The age of the natural ventila-
tion inow was computed based on the ow-eld analysis
results. For cases in which the inow temperature varied,
only the results for Cases A, C, and E are shown in Fig. 11.
The computation results were rendered non-dimensional by
the ventilation time n(reciprocal of the ventilation rate).
The indoor air age distribution responded well to ow elds.
As the inow air owed into the task zone without mixing
well with the indoor air in Case A, the age of air in the task
zone became relatively low. As the inow air mixed well
with the indoor air near the perimeter zone and owed into
the interior in Case C, the farther it owed into the interior,
the higher the age of the air became. As the inowing out-
door air owed into the interior in jets, mixing well with the
indoor air and owing deeper into the interior in Case E, a
long and narrow domain of air aged 1 or lower extended hor-
izontally to the right on the right side of the human model.
(b) Contribution ratio of the natural ventilation inlet
(SVE4). For cases in which the inow temperature varied,
only the results for Cases C–E are shown in Fig. 12. All do-
mains had 1.0 in Case A, where only natural ventilation was
0.9
>0.9
0.9 0.80.7
(a) Case C (Ar number of inflow outdoor air: -3.36)
Natural ventilation inlet
0.16m/s, 21.0°C
Perimeter zone
Ambient zone
Tas k zone
Natural venti-
lation outlet
Ceiling return
(b) Case D (Ar number of inflow outdoor air: -2.35)
22.5°C
0.8
0.9
0.8 0.8 0.7
0.7 0.6
>0.8
0.7
0.8 0.7
0.7
0.6 0.6
(c) Case E (Ar number of inflow outdoor air: -1.34)
24.0°C
Fig. 12. Range of natural ventilation inuences (vertical sections including
air conditioning air diusers).
used for the conditioning air. As the air conditioning ow
rate increased with the rise in the inow temperature, the
contribution ratio of the natural ventilation inlet decreased.
The contribution ratio of the natural ventilation inlet in the
task zone became approximately 0.7 in Case E, where the
air conditioning ow rate was the greatest.
(6) Amount of air conditioning input heat. The amounts
of air conditioning input heat when the average task zone
temperature was controlled at a constant 26◦C were obtained
as a result of micro-analysis.
(a) When the inow temperature varied. The amounts
of air conditioning input heat when the inow temperature
varied are shown in Fig. 13. The amount of air condition-
ing input heat is proportional to the natural ventilation in-
ow temperature. Each case indicated a value approximately
460 W lower than the amount of air conditioning input heat
required when an instantaneous and uniform dissipation was
assumed. This is thought to be because the natural ventila-
tion inow at a relatively low temperature did not mix well
with the indoor air, which contributed to the average tem-
perature in the task zone becoming lower than that indoors.
When the natural ventilation inow temperature was 18:8◦C
(Case A), the target air conditioning temperature in the task
zone was achieved only by natural ventilation.
(b) When the inow rate varied. The amounts of air
conditioning input heat when the inow rate varied are
shown in Fig. 14. The amount of air conditioning input heat
H. Chang et al. / Building and Environment 39 (2004) 153 – 164 163
0
200
400
600
800
1000
1200
1400
1600
17 19 21 23 25
Case A
Case A
Case B
Case C
Case D
Case E
Case B
Case C Case D
Case E
Instantaneous
and uniform dis-
sipation assumed
Task zone constant
temperature contro-
lled air conditioning
Natural ventilation inflow temperature [°C]
Amount of air conditioning
input heat [W]
Fig. 13. Natural ventilation inow temperature and amount of air condi-
tioning input heat.
Natural ventilation inflow rate (h-1)
Case F
Case G
Case C
Case H Case I
Case F
Case
Case G
Case C
Case D
Case D
Case H
Case I
1400
1200
1000
800
600
400
200
00 5 10 15 20
Amount of air conditioning
input heat [W]
Instantaneous and
uniform dissipat-
ion assumed
Task zone constant
temperature contro-
lled air conditioning
Fig. 14. Natural ventilation inow rate and amount of air conditioning
input heat.
decreased with an increase in the natural ventilation inow
rate. It indicated a value approximately 450 W lower than
the amount of air conditioning input heat when an instanta-
neous and uniform dissipation was assumed. This is thought
to be because the natural ventilation inow at temperatures
relatively lower than that indoors owed into and cooled the
task zone without mixing well with the indoor air.
(c) When the size of the inlet varied. The amounts of air
conditioning input heat when the size of the inlet varied are
shown in Fig. 15. The amount of air conditioning input heat
was inversely proportional to the height of the natural venti-
lation inlet. This is thought to be because the natural ventila-
tion inow velocity decreased with an increase in the height
of the natural ventilation inlet, causing air to enter the task
zone without mixing well with the indoor air. The amount of
air conditioning input heat in Case K became 290 W lower
than when complete indoor mixing was assumed (1/3 of the
input heat of 840 W assuming an instantaneous and uniform
dissipation; 1/6 of the total cooling load of 1700 W). Case
Amount of air conditioning
input heat [W]
Width of natural ventilation inlet [m]
Case C
Case JCase K Case C
Case J
Case K
1000
800
600
400
200
00 0.1 0.2 0.3 0.4 0.5 0.6
Instantaneous and uniform
dissipation assumed
Task zone constant
temperature controlled
air conditioning
Fig. 15. Width of natural ventilation inlet and amount of air conditioning
input heat.
30
Case C-1
Case C-2
Case C-3
Case C-4
Case C-5
Case C-6
Case C-7
Case C-2
Case C-3
Case C-1
Case C-4
Case C-5
Case C-6
Case C-7
40 50 60 70 80 90
0
500
1500
2000
2500
1000
Mechanical air conditioning only
Hybrid air conditioning
Outdoor inflow relative humidity [%]
Amount of air conditioning
input heat [W]
Fig. 16. Relationship of outdoor inow humidity and amount of air
conditioning input heat.
C indicated a value 470 W lower (1/2 of the input heat as-
suming an instantaneous and uniform dissipation; 1/4 of the
total cooling load). The height of the natural ventilation in-
let in Case C (0:5 m wide) was 5 times that of Case K, and
the amount of air conditioning input heat was as low at 2/3
that of Case K.
(d) When the inow humidity varied. The amounts of air
conditioning input heat with only mechanical air condition-
ing 5and with hybrid air conditioning 6when the natural
ventilation inow humidity varied are shown in Fig. 16. The
dehumidifying load was imposed in both cases when the in-
ow outdoor relative humidity exceeded 70%, and increased
5The amount of air conditioning input heat was assumed to be fully
mixed, and was the sum of the outdoor latent heat loads when the indoor
cooling load (1700 W + 139 W: see Table 1) and the amount of outdoor
air introduced per person in the room was 30 m3=h.
6The amount of air conditioning input heat is the sum of the amount
of sensible heat removed by the air conditioning and the amount of latent
heat obtained from the dierence in the enthalpy of the water vapour at
the air conditioning inlets and diusers.
164 H. Chang et al. / Building and Environment 39 (2004) 153 – 164
20% and 40%, respectively in Cases C-6 and C-7 with hy-
brid air conditioning compared to Cases C-1 to C-5. How-
ever, these were considerably small compared to the cases
with only mechanical air conditioning. In other words, in
these cases where the outdoor temperature was 21◦C, using
the hybrid air conditioning conserved more energy regard-
less of the outdoor humidity except for when the ventilation
openings could not be opened, such as when it was raining.
6. Summary
The eects of outdoor air conditions on hybrid air condi-
tioning utilizing natural ventilation in oce buildings were
investigated in this study. The results of this study were
derived from CFD simulation with simple geometry, there-
fore dierent results may be reached when the geometrical
layout is asymmetrical, or when the desk arrangements are
changed. The following conclusions can be drawn:
1. The higher the outdoor temperature (closer to the indoor
temperature), the lower the absolute value of the Ar num-
ber of the outdoor inow became; the natural ventilation
inow mixed better with the surrounding air, reaching
deeper into the interior, while the temperature gradient
near the oor increased.
2. The ow patterns of the natural ventilation jets varied
intricately with an increase in the inow rate, and the
high/low temperature gradient in the task-ambient zones
decreased.
3. The absolute value of the Ar number decreased with a de-
crease in the size of the natural ventilation inlet, and the
natural ventilation inow jets penetrated deep into the in-
terior while the high/low temperature gradient decreased
on the right side of the room across from the inlet.
4. Because the inow rate of the natural ventilation was
higher than that of the air conditioning air diusers, the
indoor relative humidity was inuenced by the outdoor
humidity.
5. The distribution properties of the indoor temperature and
PMV were signicantly inuenced by the temperature.
6. The indoor air age distribution properties responded well
to ow elds.
7. The contribution ratio of the natural ventilation inlet,
when the natural ventilation inow temperature varied,
decreased with the rise in the inow temperature.
8. The amount of air conditioning input heat increased with
a rise in the natural ventilation inow temperature, with
a decrease in the ow rate, and with a decrease in the
height of the natural ventilation inlet.
References
[1] Hunt GR, Linden PF. The uid mechanics of natural
ventilation-displacement ventilation by buoyancy-driven ows assisted
by wind. Building and Environment 1999;34:707–20.
[2] Lu WZ, Lo SM, Fang Z, Yuen KK. A preliminary investigation of
airow eld in designated refuge oor. Building and Environment
2001;36:219–30.
[3] Li Y, Delsante A. Natural ventilation induced by combined wind and
thermal forces. Building and Environment 2001;36:59–71.
[4] Chen ZD, Li Y. Buoyancy-driven displacement natural ventilation
in a single-zone building with three-level openings. Building and
Environment 2002;37:295–303.
[5] Eftekhari MM, Marjanovic LD, Pinnock DJ. Air ow distribution
in and around a single-sided naturally ventilated room. Building and
Environment 2003;38:389–97.
[6] Chikamoto T, Kato S, Ikaga T. Hybrid air-conditioning system in
Liberty Tower at Meiji University. Hybvent Forum 1999;99:123–7.
[7] Kato S, Murakami S, Kobayashi H. New scale for evaluating
ventilation eciency as aected by supply and exhaust openings based
on spatial distribution of contaminants. ISRACVE (Tokyo) 1992;
321–32.
[8] Song D, Kato S, Murakami S, Shiraishi Y, Chikamoto T, Nakano A.
Study on adaptive control systems under hot and humid climates—
Part 3: continuous measurement of thermal environment around a
moving person and his thermal adaptation. Summaries of Technical
Papers of the Annual Meeting of the Architectural Institute of Japan
2001;359–60 [in Japanese].
[9] Launder BE, Spalding DB. The numerical computation of turbulent
ows. Computer Methods in Applied Mechanics and Engineering
1974;3:269–89.