Building and Environment 111 (2017) 47e60

Contents lists available at ScienceDirect

Building and Environment

journal homepage: www.elsevier.com/locate/buildenv

Effect of external air-conditioner units' heat release modes and

positions on energy consumption in large public buildings
Mengtao Han a, Hong Chen b, *
a
Institute of Industrial Science, The University of Tokyo, Tokyo, Japan
b
Architecture and Urban Planning School, Huazhong University of Science and Technology, Wuhan, China

a r t i c l e i n f o a b s t r a c t

Article history: The heat released from the air-conditioners of large public buildings, which has long-term and far-
Received 14 June 2016 reaching influences, not only changes the thermal environment surrounding the buildings but also
Received in revised form profoundly impacts the air-conditioner energy consumption. In this paper, a quantitative evaluation of
24 October 2016
the impact of the heat release modes and positions of air-conditioner external units on both the envi-
Accepted 25 October 2016
Available online 26 October 2016
ronment surrounding the buildings and the air-conditioner energy consumption of large public buildings
is conducted using a coupled simulation of convection, radiation and conduction and air-conditioner
energy cost. Principles for external unit positions are explored as well.
Keywords:
Large public buildings
© 2016 Elsevier Ltd. All rights reserved.
Air-conditioner energy consumption
Heat release
External air-conditioner units

1. Introduction
Accompanied by the explosion of urbanization, the heat island
phenomenon in cities has been seriously aggravated with wors-
ening thermal environments [1]. (Fig. 1).
To relieve this situation, recent studies have simulated air-
conditioner energy consumption to reduce this consumption and
to increase air-conditioner efficiency [2e4]. Research on the effect
of air-conditioner heat release on the outside thermal environment
is already well known [5]. In contrast, less attention has been paid
to the impact of heat release modes and external air-conditioning
unit positions on air-conditioner energy consumption.
The heat release not only changes the thermal environment
surrounding large public buildings but also impacts the air-
conditioner energy consumption of the buildings due to the large
and long-term heat release in dense cities. Some researchers have
discovered that the air-conditioner cooling energy consumption
increases when the outdoor temperature increases due to the air-
conditioner heat release [6,7]. Moreover, other researchers have
indicated that varying the positions of external air-conditioner
units can affect the environment; for example, lower external
units will affect the environment, especially the areas of human
* Corresponding author.
E-mail address: chhwh@hust.edu.cn (H. Chen).
activities, more seriously than higher external units on the roof
[8,9].
Unfortunately, external air-conditioner units are still placed
merely in compliance with the visual aesthetics of buildings. Very
little has been achieved in implementing systematic research of the
relationship between the external units and air-conditioner energy
consumption.
The impact of various heat release modes and external unit
positions on the air-conditioner energy consumption is analysed in
this study using a coupled simulation of convection, radiation and
conduction and energy cost. A quantitative evaluation is advocated
with principles for exploring the external unit positions. This is a
beneficial attempt to reduce the air-conditioner energy consump-
tion of large public buildings and to improve the thermal envi-
ronment outside.
2. Research outline
2.1. Method
Due to the variation of the air temperature, humidity and ve-
locity around buildings caused by air-conditioner heat release,
three main methods of energy exchange between the inner build-
ing and outer environment have changed, altering both the sensible
and latent heat energy consumption (cooling and dehumidifica-
tion) of air-conditioners. These three methods are heat exchange

http://dx.doi.org/10.1016/j.buildenv.2016.10.014
0360-1323/© 2016 Elsevier Ltd. All rights reserved.
48 M. Han, H. Chen / Building and Environment 111 (2017) 47e60

Fig. 1. Air-conditioner heat release intensifying the heat island phenomenon (Top: a picture of air-conditioner external units; bottom: the heat release shown by an infrared
camera).

through the building envelope, heat exchange caused by necessary
ventilation and heat exchange caused by infiltration due to poor
airtightness of the envelope. Some studies indicate that the heat
exchange through a building envelope has much less of an impact
than the other two ways, approximately 1/400 of the other two
[11]. Therefore, only heat exchanges by ventilation and infiltration
are considered in this research.
Fig. 3 shows the technical route of this research: (1) the initial
air-conditioner energy consumption and heat release are simulated
based on the typical local weather condition; (2) the air velocity,
temperature and relative humidity around the building and the
average temperature and humidity of the layer adjacent to the
building are simulated using a coupled simulation of convection,
radiation and conduction with the heat release calculated in step 1
used as the boundary conditions; (3) the new energy consumption
is calculated by simulating the air-conditioner energy consumption
as affected by the heat release according to step 2.
Two kinds of computer simulation is used in this research,
which are both already been validated and applied. The core
simulation of this research is called a coupled simulation of con-
vection, radiation and conduction. This method is developed by
Chen and Ooka [10e12] and has been successful validated and
applied in several researches [10] for over ten years. The simulation
method of air-conditioner energy consumption and heat release
using DeST-C is developed by Tsinghua University, which also
successfully has been applied in several researches for almost

Fig. 2. Heat balance of the building.

 M. Han, H. Chen / Building and Environment 111 (2017) 47e60
Fig. 3. Technical route of the research.
twenty years [13,14].
The air-conditioner energy consumption of a building is calcu-
lated according to Equation (1):
Qsun rad þ Qcond þ Qroom rad þ Qven þ Qinf þ Qocu þ Qhvac
¼0
where Qsun rad the solar radiation heat gain, Qcond is the heat con-
duction through the building envelope, Qroom rad is the radiation
heat transfer from the surfaces in the room, Qven is the heat lost
through ventilation, Qinf is the heat lost through infiltration, Qocu is
the heat gain from the people and equipment in the building, and
Qhvac cooling is the air-conditioner internal heat load, which is usu-
ally negative for cooling in summer and is released outside by the
external units, part of which then affects the ventilation and infil-
tration loads. The heat balance in the building is shown in Fig. 2.
In this research, Qocu, Qroom rad and Qsun rad are constant because
all the scenarios have the same architecture and weather condi-
tions. In addition, the variation of the air-conditioner load caused
by the changes in heat conduction DQcond can be neglected [11].
Therefore, the factors that affect the air-conditioner energy con-
sumption and heat release are simplified to the heat loss caused by
ventilation and infiltration, as shown in Equation (2):
DQair ¼ DQven þ DQinf
Furthermore, both sensible and latent heat losses occur through
air exchange by ventilation and infiltration, as shown in Equation
(3):
 and a split-type air-conditioner system. The central air-conditioner


DQair ¼ DQven sinsible þ DQven latent þ DQinf sinsible
 cooled condenser (VRV); the split-type air-conditioner systems
þ DQinf latent
Equation (3) indicates that the air-conditioner heat load is
determined by only the variation of the sensible and latent heat
caused by both ventilation and infiltration from the building en-
velope, which is the basis of this research.
49
2.2. Case study
The block model is shown in Fig. 4. A continuous block filled
with the same high-rise office is adopted for all cases. The high-rise
cooling is 24 floors tall with a total height of 85.2 m and a construction area
of 78630 m2. The bottom 5 floors are used as the annex for large
offices, with a floor height of 4.5 m and a total area of 32650 m2. The
(1) high-rise part contains 19 standard floors for standard small offices,
with a floor height of 3.3 m and a total area of 45980 m2. Within the
block, the building density is 38.7%, and the capacity rate is 4.68.
The rate of greenbelt around the buildings is 28.9%.
Solar radiation simulation, covering a continuous 48 h from 0:00
August 13th to 24:00 August 14th (in the middle of summer), is
conducted to obtain the initial radiation data. Considering the
thermal storage of the building envelope, the result of 14:00 on
August 14th is used as the boundary condition for the steady-state
CFD simulation. As shown in Fig. 4. The middle building in the
seventh row becomes the target in a continuous block so that the
canyon's influence could be considered adequately. And the pre-
liminary simulation shows that the turbulence becomes stable
starting from the 5th row. The block size is 410.4 m
(width)  780.4 m (length)  85.2 m (height).
An unstructured grid system, as shown in Fig. 5, is used for both
the CFD and solar radiation simulations. A rectangular velocity field
is covered by a hexahedral grid space. The length of the fluid field is
(2) 4H þ Block field þ10H; the width is 5H þ Block fieldþ5H; and the
height is 6H, where H is building's height. The total grid count is 1.4
million.
Three typical types of heat release modes are simulated after a
survey and divided into two types: a central air-conditioner system
system contains two modes: a cooling tower condenser and an air-
(3) contains only an air-cooled condenser. For the central systems, a
total of 15 cases are set up according to the relationship between
the external unit positions and the buildings with two different
wind directions. These cases are shown in Table 1, and a diagram of
the positions is shown in Fig. 6; “Case-Up” is a reference case
50 M. Han, H. Chen / Building and Environment 111 (2017) 47e60

Fig. 4. Block model for research.

Fig. 5. Unstructured grid system used in the CFD simulation.

without heat release. In addition, considering the various positions
of the external units, the split-type system cases are individually set
up, as shown in Table 2 and the diagram is shown in Fig. 7.
2.3. Boundary conditions
2.3.1. Boundary conditions for the simulation
A two-day solar radiation simulation is carried out. The results
of 14:00 on the second day is adopted for part of the initial CFD
condition. The parameters for the initial solar radiation simulation
are shown in Table 3, and the CFD parameters are shown in Table 4.
The solar radiation background parameters vary over time, so
Table 3 shows 14:00 on the second day as an example.
2.3.2. Validation of the boundary conditions
In this research, the background data used for the solar radiation
simulation are from a field measurement in Wuhan by our research
team lasting for two years that was carried out for two days (July
31-August 1) every summer [15e17]. Data of the hours from 9:00 to
17:00 were adopted. Approximately 10 fixed points are spread out
along the streets in the block besides the Yangtze River, which is the
largest river in China; Fig. 8 shows the sites.
Among the measurement results, the background is used in this
study, as shown as Fig. 9, which characteristically describes Wuhan
City's weather condition [17].
On the other side, the boundary conditions for the air-
conditioner energy consumption simulation, and the parameters
of the building envelope's thermal performance and the air-
 M. Han, H. Chen / Building and Environment 111 (2017) 47e60 51

Table 1
Analysis cases for the central air-conditioner system.

Case groups Position of the external units Heat release mode Case name

Location Wind direction

Referenced case e e no heat release Case-Up

Case-G-U Ground Windward VRV Case-G-U-F

cooling tower Case-G-U-V
Case-G-D Downwind VRV Case-G-D-F
cooling tower Case-G-D-V
Case-G-P Parallel to wind VRV Case-G-P-F
cooling tower Case-G-P-V

Case-A-U Top of the annex Windward VRV Case-A-U-F

cooling tower Case-A-U-V
Case-A-D Downwind VRV Case-A-D-F
cooling tower Case-A-D-V
Case-A-P Parallel to wind VRV Case-A-P-F
cooling tower Case-A-P-V

Case-R Top of the roof e VRV Case-R-F

cooling tower Case-R-V

Fig. 6. Diagram of the different positions of the external central air-conditioner units.

conditioner services are quoted from a Chinese official energy ef-
ficiency architecture design standard [18].
Third, the parameters of the outlet of the air-conditioning sys-
tems calculated from the heat release refer to the research by Chen
Hong, OOKA [12] and real commercial air-conditioner systems. To
validate the outlet parameters, another air-conditioner field
measurement was carried out in an office high-rise in Wuhan city
in July 2012, in which the three types of air-conditioner heat release
modes were all used: the cooling tower air-conditioner, a few small
office VRV systems and split-type air-conditioners. The general
measurement situation is shown in Table 5 and Fig. 10.
Figs. 11 and 12 show the measurement results. The tendency of

Table 2
Analysis cases for the split-type air-conditioner system.

Case Name Position of the external units

Location Wind direction

Split-U-S Outside surface of the building Windward

Split-D-S Downwind
Split-P-S Parallel to wind

Split-U-O In the gaps of the building, exhausting out of gaps Windward

Split-D-O Downwind
Split-P-O Parallel to wind

Split-U-E In the gaps of the building, exhausting to gaps Windward

Split-D-E Downwind
Split-P-E Parallel to wind

Split-R Top of the roof e

52 M. Han, H. Chen / Building and Environment 111 (2017) 47e60
Fig. 7. Diagram of the different positions of the external split-type air-conditioner units.
Table 3
Boundary condition for the solar radiation simulation (14:00, 2nd day).
Air temperature 34.9  C Relative humidity
Solar altitude angle 102.3 Solar azimuth angle
Solar radiation 816 W/m2
the outlet temperature of the three heat release modes is obvious,
following the trend of the outdoor air temperature. The cooling
tower outlet temperature is 1  C above the outdoor temperature on
average, while the VRV outlet temperature is 8e10  C higher and
the split-type outlet temperature is 9  C higher. The outlet velocity
is calculated according to the heat release, outlet temperature and
outlet area in this research. The field measurement results show
that the outlet velocity fluctuates widely. The cooling tower outlet
velocity is approximately 8e18 m/s, while that of the VRV is
4e10 m/s and that of the split-type is relatively constant at 6e7 m/
s. In conclusion, the parameters of the air-conditioner outlet in this
study correlate well with the field measurement results.
2.4. Evaluation index
The mean temperature, humidity and velocity of a cubic volume
around the target (Tvol,RHvol,Vvol) are discussed in common studies
to describe the thermal environment around a building, which is a
simple and convenient method but with great deviation. In addi-
tion to these indexes, several new and more accurate indexes are
discussed in this research. Usually, ventilation and infiltration goes
through windows and doors of the building envelope. Therefore,
the key evaluation indexes of this research are the average air
temperature, humidity and velocity of the grid layer adjacent to the
building surface (TNS,RHNS,VNS), which is the direct source of the
building's ventilation and infiltration. This layer is approximately
1.5 m thick in the grid system used in this research.
3. Impact of heat release on the thermal environment around
buildings
54.0%
106.0 3.1. Central air-conditioner system
Part of the CFD results is shown in Figs. 13 and 14, showing how
different heat release modes and positions affect the environment.
Fig. 13 shows that the moisture released by a cooling tower will
increase the humidity of the atmosphere; the lower the external
units are located, the more humid the atmosphere will be.
Fig. 14 shows that the heat released by VRV will disperse over
the whole block when the external units are located parallel to the
wind, heating the entire block, compared with the other situations.
The section pictures further show that the released heat will sur-
round the building when the external units are located in the
shadow areas of the wind.
The indexes of Tvol,Vvol and TNS,VNS are discussed to describe the
impact of the central air-conditioner heat release on the thermal
environment and the layer adjacent to the building envelope.
Fig. 15 shows the difference between the mean air temperature
and velocity (Tvol,Vvol) in the cubic space around the target build-
ing: (1) Vvol of the cooling tower air-conditioner is higher than Vvol
with a VRV; (2) Tvol of the cooling tower air-conditioner is lower
than Tvol of the VRV; and (3) a higher Vvol usually has a lower Tvol.
TNS of the VRV is shown in Table 6. Different from Tvol, the
maximum value of TNS in Case-G-D-F is 37.025  C because a huge
heat release accumulates in the vortex area at the back of the
building, leading to a difficulty in spreading to other areas and
resulting an obvious increase of air temperature adjacent to the
building envelope. On the other hand, TNS in Case-G-P-F also in-
creases obviously; the heat release on the lateral side of the
building is “blown” to the vortex area behind the building by the
oncoming wind, causing an increase of the air temperature in the
street canyon.
Moisture release affects the air-conditioner energy consumption
 M. Han, H. Chen / Building and Environment 111 (2017) 47e60 53

Table 4
Boundary conditions for CFD.

Turbulence model Steady standard kε model (including the buoyancy effect)

Momentum First order upwind

Turbulent kinetic energy First order upwind

Turbulent dissipation energy First order upwind

Solver algorithm Semi-Implicit method for pressure-linked equations (SIMPLE)

Inlet
 1
Z 4
U ¼ U0 (4)
Z0

U0 ¼ 2:0m=s，Z0 ¼ 10:0m

k ¼ 1:5ðI  U0 Þ2 ，I ¼ 0:1 (5)

3 3
Cm k2
1 Z Z4
ε¼ ，l ¼ 4 Cm k 2 0 (6)
l U0

Sky Free slip

Side Symmetry

Wall Generalized logarithmic law

Outlet of the central air-conditioning unit (VRV) Total outlet Area: A ¼ 30 m2

Velocity: V ¼ 8.33 m/s
Temperature:T ¼ 45  C (10  C higher than the atmospheric temperature)
Relative humidity: RH ¼ 54% (the same as the atmosphere)
Density: rair ¼ 1.165 kg/m3
k ¼ 0.1 m2/s2, ε ¼ 1 m2/s3

Outlet of the central air-conditioning (cooling tower) Total outlet area: A ¼ 30 m2

(In the cooling tower, 90% of the heat release is exhausted Velocity: V ¼ 25.7 m/s
in the form of latent heat, and the other 10% is sensible heat. [12]) Temperature:T ¼ 36  C (1 Chigher than the atmospheric temperature)
Relative humidity: 65.2% (11.2% higher than the atmospheric humidity)
Density: rair ¼ 1.165 kg/m3
k ¼ 0.1 m2/s2, ε ¼ 1 m2/s3

Outlet of the split-type air-conditioning unit Velocity: V ¼ 5.9 m/s

Temperature:T ¼ 44  C (9  C higher than the atmospheric temperature)
Relative humidity: RH ¼ 54% (the same as the atmosphere)
Density: rair ¼ 1.165 kg/m3
Outlet area of one external unit: A0 ¼ 0.66 m2
k ¼ 0.1 m2/s2, ε ¼ 1 m2/s3

Fig. 8. Measurement sites and points distribution.

through dehumidification and air temperature. Fig. 16 shows the three cases in which the heat release is located on the ground
impact of moisture release on the moisture content of the layer (Case-G-“X”) is much more serious than that of the three cases
adjacent to the building envelope. The moisture variation of the where the heat release is located on the roof of the annex building
54 M. Han, H. Chen / Building and Environment 111 (2017) 47e60

Fig. 9. Wind rose (left) and temperature and humidity (right) results of the background data.

Table 5
Air-conditioner field measurement situation.

Acquisition item point Acquisition item Time and interval Acquisition position

Background weather Temperature, humidity, radiation, velocity 0:00e24:00, Interval of 2 min 6 m above the highest roof of the building

Cooling tower Outlet temperature, outlet humidity 0:00e24:00, Interval of 2 min Vertically to the outlet of the cooling tower
Outlet velocity 0:00e24:00, Interval of 6 min

VRV Outlet Temperature, Outlet Humidity 0:00e24:00, Interval of 2 min Vertically to the outlet of the VRV
Outlet velocity 0:00e24:00, Interval of 6 min

Split-type Outlet temperature, outlet humidity 0:00e24:00, Interval of 2 min Vertically to the outlet of the split-type air-conditioner
Outlet velocity 0:00e24:00, Interval of 6 min

Fig. 10. Field measurements of the air-conditioner outlets.

Fig. 11. Outlet temperature of the various air-conditioner systems.

 M. Han, H. Chen / Building and Environment 111 (2017) 47e60 55

Fig. 12. Outlet velocity of the various air-conditioner systems.

Fig. 13. Environmental moisture content affected by different external unit positions for the cooling tower cases (Top: the 22.5-m-high plan, bottom: the section).

(Case-A-“X”). On the other hand, there is almost no moisture
change in the case where the heat release is located on the roof of
the high-rise (Case-R-“X”) because the heat is blown into the high
atmosphere.
3.2. Split-type air-conditioner systems
Fig. 17 shows that the spilt-type external units' influence TNS
much more severely compared with the central air-conditioner
systems due to the large number of external units and their uni-
form distribution on the façade with high outlet temperatures.
Enormous difference is also shown among the split-type cases in
Fig. 17. The increase in air temperature is the greatest when the
outlets are located in the gaps of the building and blow into the
gaps (Split-“X”-E) because of the high accumulation of heat in the
gaps that are less influenced by airflow.
Fig. 18 shows the relationship between the heat release loca-
tions and the temperature addition of the layer adjacent to the
envelope (TNS): (1) If only the relationship between the release
locations and the wind direction is considered, the trend of the TNS
addition is downwind > windward > parallel to wind (shown as the
height of the three curves in Fig. 18). (2) If only the relationship
between the heat release locations and the building is considered,
the tendency of the TNS addition is location in gaps and outlet into
gaps > location in gaps and outlet leaving gaps > location outside of
gaps and outlet to environment (shown as the tendency of each
curve in Fig. 18).
4. Variation of energy consumption of the air-conditioner
4.1. Central air-conditioner systems
4.1.1. Variation of energy consumption caused by sensible heat
release
Fig. 19 shows the central air-conditioner energy consumption
variation caused by air temperature change with heat release
56 M. Han, H. Chen / Building and Environment 111 (2017) 47e60

Fig. 14. Environmental temperature affected by different external unit positions for the VRV cases (Top: the 1.5-m-high plan, bottom: the section).

Fig. 15. Average temperature and average wind velocity in the volume around the building.

Table 6 energy consumption when using a VRV is certainly higher than that
TNS of the VRV cases. when using a cooling tower in all of the cases. This is because
Case name TNS( C) Increase ( C) sensible heat is the main form of VRV heat release, which
tremendously increases the surrounding air temperature, espe-
Case-G-U-F 36.447 0.093
Case-G-D-F 37.025 0.671
cially adjacent to the building envelope, causing energy consump-
Case-G-P-F 36.728 0.375 tion to increase. Nevertheless, latent heat and very little sensible
Case-A-U-F 36.316 0.037 heat with high-speed wind is the main form of the cooling tower's
Case-A-D-F 36.462 0.142 heat release, so that the airflow in the street canyon increases and
Case-A-P-F 36.379 0.026
the temperature decreases around the building, decreasing the air-
Case-R-F 36.328 0.026
conditioner energy consumption.

compared with the referenced case of Case-Up. The diagram shows 4.1.2. Variation of energy consumption caused by latent heat
that the energy consumption increases to varying degrees when release
VRV external units are used, while all cases decrease to varying Table 7 shows the variation of energy consumption caused by
degrees when using cooling tower external units. The absolute latent heat release, which differs significantly from that caused by
 M. Han, H. Chen / Building and Environment 111 (2017) 47e60 57

Fig. 16. Impact of the heat release location on the moisture content of the air layer adjacent to the building.

Fig. 17. Addition of TNS in the split-type air-conditioner cases.

Fig. 18. Relationship between the addition of TNS and the external unit position in the split-type air-conditioner cases.

sensible heat release (Fig. 19). Due to increased humidity around
the building, the cooling tower increases the energy consumption
for dehumidification, which is considerably more than that caused
by sensible heat release. A large amount of humid air from the
cooling tower accumulates around the building, tremendously
increasing the indoor humidity through ventilation and infiltration
and increasing the air-conditioner energy consumption for dehu-
midification. When VRV is used, the latent heat does not increase
due to their fan-cooled heat release without moisture.
4.1.3. Total variation of air-conditioner energy consumption
Fig. 20 shows the total energy consumption addition in all
central air-conditioner case groups caused by varying heat release
mode. The total energy addition of the cases of the cooling tower is
much more than that of the cases that use VRV. Fig. 21 shows the
addition by heat release mode and concludes the following: (1)
Sensible heat addition, including ventilation and infiltration, is the
main method for increasing the air-conditioner energy consump-
tion in the VRV cases, while dehumidification is the main method
for the cooling tower cases. (2) If only the sensible heat release is
considered, including ventilation and infiltration, there are both
increases and decreases of energy consumption in the VRV cases,
mainly increasing; the cooling tower cases only decrease due to the
high-speed, high-humidity and low-temperature outlet conditions.
(3) The additional energy consumption for dehumidification of the
cooling tower cases increase dramatically, even more than the
decrease in energy caused by the air temperature decreasing
around the building, resulting in an overall severe energy con-
sumption increase, greater than the increases in the VRV cases. (4)
The minimum increase (in Case-A-U-F) is one-sixth the maximum
58 M. Han, H. Chen / Building and Environment 111 (2017) 47e60

Fig. 19. Air-conditioner energy consumption changes caused by sensible heat release (Case-Up's energy consumption change is 0).

Table 7
Air-conditioner energy consumption changes caused by latent heat release.

Case VRV (-F) Cooling tower (-V)

Groups
Latent energy consumption addition for The incremental Latent energy consumption addition for The incremental
dehumidification (kW) percentage dehumidification (kW) percentage

Case-G-U 0 0% 90.109 þ7.67%

Case-G-D 96.998 þ8.26%
Case-G-P 86.395 þ7.36%
Case-A-U 67.246 þ5.73%
Case-A-D 0.761 þ0.06%
Case-A-P 81.384 þ6.94%
Case-R 60.490 þ5.15%

Fig. 20. Central air-conditioner energy consumption changes due to different external unit positions.

increase (in Case-G-D-V), indicating the significant effect of than that of the central air-conditioner cases in terms of sensible
different heat release modes and external unit positions of central heat. In particular, the increase is the highest when the heat release
air-conditioner systems on the air-conditioner energy location is windward or downward of the building because of the
consumption. heat accumulation adjacent to the building envelope, for which
ventilation contributes the most. In contrast, the energy con-
sumption increase is smaller when the heat release location is
4.2. Split-type air-conditioner systems
lateral to the building or parallel to the wind and increases the least
when the heat release location is on the roof of the high-rise.
Table 8 shows the energy consumption addition of the split-type
air-conditioner cases. It is obvious that the increase is much higher
 M. Han, H. Chen / Building and Environment 111 (2017) 47e60 59

Fig. 21. Increased energy consumption of central air-conditioner cases.

Table 8
Split-type air-conditioner energy consumption changes.

Cases Energy consumption addition from ventilation (kW) Energy consumption addition from infiltration (kW) Total increase in energy consumption (kW)

Split-U-S 111.525 16.367 127.892

Split-D-S 132.179 19.749 151.927
Split-P-S 11.859 1.638 13.498
Split-U-O 221.810 32.947 254.757
Split-D-O 230.303 34.202 264.505
Split-P-O 23.525 3.279 26.804
Split-U-E 241.557 35.559 277.117
Split-D-E 289.445 42.715 332.160
Split-P-E 55.316 7.409 62.725
Split-R 30.054 6.425 36.479

5. Conclusions
In this research, a coupled simulation of convection, radiation
and conduction and the energy cost of air-conditioner is used to
discuss the impact of different heat release modes, including a
central air-conditioner (cooling tower and VRV) and a split-type air-
conditioner, and various external unit positions on air-conditioner
energy consumption in large public buildings.
(1) A split-type air-conditioner will affect the environment
around the building more seriously but in a smaller area
compared with a central air-conditioner.
(2) Fan-cooled sensible heat release, including VRV and split-
type, will increase the air temperature around building,
usually increasing the air-conditioner energy consumption.
(3) Due to latent heat release, a cooling tower air-conditioner
will tremendously increase the humidity around the build-
ing, enormously increasing the energy consumption for
dehumidification but with a slight decrease in sensible
cooling energy consumption.
(4) The use of a cooling tower air-conditioner will result in much
more energy consumption than that of a VRV if sensible and
latent heat release are both considered. Otherwise, the
opposite is true if only sensible heat release is considered.
60 M. Han, H. Chen / Building and Environment 111 (2017) 47e60
(5) Different positions of external units will lead to different
energy consumption even using the same heat release mode.
(6) If only wind direction is considered, the energy consumption
tendency of both central and split-type air-conditioners is as
follows: downwind > windward > parallel to wind. When
external units are located downwind or windward, the en-
ergy addition increases much more for the aggravation of
heat release around the building envelope, especially
through ventilation. When located parallel to the wind, the
total energy addition increases less and addition by infiltra-
tion dominates. However, the impact on the outside envi-
ronment is more serious.
(7) In the central air-conditioner cases, locating external units on
the ground and at the back of the building forms a vortex
zone that will lead to the greatest increase in energy con-
sumption, while locating them on the roof will lead to the
smallest increase. In general, if only the positions are
considered, the trend of the energy consumption increase is
as follows: located on the ground > located on top of the
annex > located on top of the high-rise roof. Therefore, the
higher the external units are located, the smaller the air-
conditioner energy consumption increase is.
(8) In the split-type air-conditioner cases, the tendency of the
energy consumption increase by positions is as follows:
located in gaps with the outlet into the gaps > located in gaps
with the outlet facing out of the gaps > located outside of the
gaps with the outlet to the environment. The energy con-
sumption will be more seriously affected when the external
units are located in the gaps of the building because of heat
accumulation in the gaps.
In conclusion, the heat release modes and positions of the
external units tremendously influence air-conditioner energy
consumption, so they should be considered with care in the ar-
chitecture design process.
Acknowledgement
This paper was supported by projects of the National Natural
Science Foundation of China Project No. 51538004.
References
[1] H. Akbari, K. Hanaki, Calculating energy-saving potentials of heat-island
reduction strategies, Energy Policy 33 (2007b) 721e756.
[2] J. Tanimoto, A. Hagishima, P. Chimklai, An approach for coupled simulation of
building thermal effects and urban climatology, Energy Build. 36 (2004)
781e793.
[3] J. Tanimoto, A. Hagishima, H. Sagara, A methodology for peak energy
requirement considering actual variation of occupants' behaviour schedules,
Build. Environ. 43 (2008) 610e619.
[4] T. Ihara, Y. Genchi, T. Sato, K. Yamaguchi, Y. Endo, City-block-scale sensitivity
of electricity consumption to air temperature and air humidity in business
districts of Tokyo, Japan, Energy Build. 33 (2008) 1634e1645.
[5] M. Bojic, M. Lee, F. Yik, Influence of a depth of a recessed space to flow due to
air-conditioner heat rejection, Energy Build. 34 (2002) 33e43.
[6] C.M. Hsieh, T. Aramaki, K. Hanaki, The feedback of heat rejection to air con-
ditioning load during the night in subtropical climate, Energy Build. 39
(2007a) 1175e1182.
[7] C.M. Hsieh, T. Aramaki, K. Hanaki, Estimation of heat rejection based on the air
conditioner use time and its mitigation from buildings in Taipei City, Build.
Environ. 42 (9) (2007b) 3125e3137.
[8] S. Dhakal, K. Hanaki, Improvement of urban thermal environment by man-
aging heat discharge sources and surface modification in Tokyo, Energy Build.
34 (2002) 13e23.
[9] R. de Dear, J. Richard, G.S. Brager, Thermal comfort in naturally ventilated
buildings: revisions to ASHRAE Standard 55, Energy Build. 34 (6) (2002)
549e561.
[10] Hong Chen, Ryozo Ooka, Study on outdoor thermal environment of apartment
block in Shenzhen, China with coupled simulation of convection, radiation
and conduction, Energy Build. 36 (2004) 1247e1258.
[11] H. Chen, R. Ooka, H. Huang, M. Nakashima, Study on the impact of sensible
heat flux from building walls and artificial heat release from buildings on
outdoor thermal environments using coupled simulations of convection, ra-
diation, and conduction, in: The Six International Conference on Urban Cli-
mate(ICUC6), 2006.
[12] H. Chen, R. Ooka, H. Huang, et al., Study on mitigation measures for outdoor
thermal environment on present urban blocks in Tokyo using coupled
simulation, Build. Environ. 44 (11) (2009) 2290e2299.
[13] Tianzhen Hong, Yi Jiang, A new multizone model for the simulation of
building thermal performance, Build. Environ. 32 (No.2) (1997) 123e128.
[14] D.P. Bloomfield, Y. Jiang, A program documentation system and the work of
IEA annex 31, in: Proc. Systems Simulation Conference, University of Liege,
December 1990.
[15] M. Han, H. Chen, L. Yuan, et al., CFD analysis on cooling effect on a complex
traditional urban area with river wind in summer, in: International Confer-
ence on Multimedia Technology(ICMT), Hangzhou, China, 2011.
[16] Y. Cai, H. Chen, L. Yuan, et al., Study of comparison between effects on micro-
climate of two sides of Yangtze River with river wind in Wuhan in winter,
Adv. Mater. Res. 361e363 (2011) 1177e1181.
[17] G. Han, H. Chen, L. Yuan, et al., Field measurements on micro-climate and
cooling effect of river wind on urban blocks in Wuhan city, in: International
Conference on Multimedia Technology (ICMT), Hangzhou, China, 2011.
[18] Ministry of Housing and Urban-Rural Development of the People’s Republic of
China, Design Standard for Energy Efficiency of Public Buildings (GB50189-
2005), China Architecture and Building Press.