Accepted Manuscript

On the effect of provision of balconies on natural ventilation and thermal comfort in

high-rise residential buildings

Sara Omrani, Veronica Garcia-hansen, Bianca Capra, Robin Drogemuller

PII: S0360-1323(17)30302-5
DOI: 10.1016/j.buildenv.2017.07.016
Reference: BAE 4995

To appear in: Building and Environment

Received Date: 22 April 2017

Revised Date: 19 June 2017
Accepted Date: 11 July 2017

Please cite this article as: Omrani S, Garcia-hansen V, Capra B, Drogemuller R, On the effect of
provision of balconies on natural ventilation and thermal comfort in high-rise residential buildings,
Building and Environment (2017), doi: 10.1016/j.buildenv.2017.07.016.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

On the effect of provision of balconies on natural ventilation and thermal comfort in high-

rise residential buildings

Abstract

PT
Natural ventilation and balconies are two of the most desirable features of a living space in

subtropical climates. The aim of this paper is to investigate the effect of balconies on natural

RI
ventilation performance and thermal comfort of residential buildings. To this end, in-situ full-scale

measurements were carried out for Computational Fluid Dynamics (CFD) model validation and

SC
further analysis. A number of parameters such as balcony type, balcony depth, ventilation mode,

U
and wind angle were used in developing case studies. Once validated, the CFD model was used for

investigation of air movement inside each case study. Combined and separate effects of the defined
AN
parameters on natural ventilation performance were evaluated using air velocity and Standard

Effective Temperature (SET*) as criteria. The results indicate that the addition of a balcony to a
M

building with single-sided ventilation can improve the ventilation performance. In contrast, indoor
D

air velocity was reduced as a result of balcony addition when the case study was operated in cross
TE

ventilation mode. Furthermore, ventilation performance of single-sided ventilation was found to be

more sensitive to the change of parameters compared to that of the cross ventilation. It has also
EP

been found that among the investigated parameters, incident wind angle affects the ventilation

performance most for both natural ventilation modes.

C

Keywords: Natural ventilation; CFD; balcony; thermal comfort; single-sided ventilation; cross
AC

ventilation

1- Introduction

Natural ventilation is proven to be an effective low-cost solution for space conditioning, especially

in cooling dominant climates [1, 2]. Being a passive solution, building energy consumption and

Page 1
 ACCEPTED MANUSCRIPT

associated negative environmental effects can be reduced by implementation of natural ventilation.

Furthermore, building occupants in subtropical climates have a tendency to live in naturally

ventilated buildings rather than fully air-conditioned spaces [3].

In addition to the external weather conditions as the main driving force, architectural design

PT
features play an important role in natural ventilation performance and indoor airflow behaviour.

RI
Design parameters that alter the internal airflow include type, size and placement of the openings,

internal layout, height and orientation of a building, and façade features such as balconies [4-8].

SC
Private outdoor spaces such as balconies are perceived as one the most desired features in

U
subtropical climates that can be used for a different range of activities [3, 9, 10]. Balconies act as

buffer spaces between indoor and outdoor that not only reduce the occupants’ exposure to the
AN
pollutions [11] but also result in significant heating and cooling load reduction [12]. In addition,

balconies can reduce noise level –the commonly stated limitation of natural ventilation- by acting as
M

an acoustic protection device [13].

D

From a natural ventilation point of view, the addition of a balcony alters the pressure distribution
TE

on a building façade and consequently affects the ventilative forces [14]. Chand et al. [14] carried

out a wind tunnel experiment on a five-storey building with mounted balconies to study this
EP

impact. Their results demonstrated an alteration in pressure distribution on the windward side and

no significant change on the leeward side. While Chand et al’s study focused on pressure
C

distribution on the façade of a case model without openings, their experimental data was later used
AC

for CFD validation and subsequent evaluation of the effect of balcony provision on indoor

ventilation performance [15], and thermal comfort [16]. The results indicated that mass flow rate

increases and average velocity decreases in the case of single-sided ventilation, while no significant

change was observed under cross ventilation mode [15]. Thermal comfort status was also reported

with no change [16]. Prianto and Depecker [17, 18] adopted a numerical method to investigate the

Page 2
 ACCEPTED MANUSCRIPT

effect of balcony, internal divisions, and openings on indoor velocity and thermal comfort in a two-

storey dwelling. They found that both balconies and openings play an important role in the

modification of indoor velocity and thermal comfort condition.

While these studies have been concerned with the effect of balconies on natural ventilation, they

PT
were all based on simple geometries, and the combined effect of balcony features (i.e. balcony type

RI
and depth) with other determinant parameters such as ventilation mode and incident wind

direction are not adequately investigated. The objective of this study, therefore, is to investigate the

SC
impact of these parameters on natural ventilation and indoor thermal conditions. Accordingly, full-

scale measurements were carried out in a residential unit located in a high-rise residential building

U
in Brisbane, Australia. The collected data was then used for validation of a CFD model and good
AN
agreement between the data and the simulation results were obtained. Two ventilation modes

(single-sided and cross ventilation), two balcony types (semi-enclosed and open balcony), four
M

balcony depths (10%, 20%, 30%, and 40%), and four wind directions (0˚, 45˚, 90˚, and 180˚) were

defined as variables. From that 70 case studies were formulated to investigate the separate and
D

combined effect of these variables. The validated CFD model was then used for calculation of air
TE

velocity in the case studies.

EP

Average velocity was used as a criterion to evaluate the effect of the variables on overall ventilation

performance. Average velocity is linearly correlated with qualities such as airflow rate and air
C

change per hour [19], and is also a determinant in thermal comfort calculations. Therefore, average
AC

velocity can be used as a good indicator of ventilation performance. Acquired average velocity along

with typical meteorological data for Brisbane were further used in calculations of SET* index for

thermal comfort evaluation of the occupied zone.

Page 3
 ACCEPTED MANUSCRIPT

2- Method of analysis

2-1- Field measurement

Field measurements were carried out in a unit located on the fifth floor of a 36-storey residential

PT
building located at Brisbane Central Business District (CBD), Australia. The building is oriented 35°

toward the west and the case study unit is located at the eastern end of the building. Figure 1 shows

RI
the case study building and its surroundings where the case study building is indicated in red

boundary.

U SC
AN
M
D
TE
C EP

Figure 1: Case study building (right) and case study surroundings (left). The case study building and the case
AC

study unit are indicated with red boundary.

The case study layout consists of two balconies at two opposite sides of the living area which

allowed measurements for both single-sided and cross ventilation configurations (Figure 2).

Balcony doors were identical with the operable area of 1.16m x 2.5m=2.9m2.

Page 4
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 2: Case study building plan layout [20].

Air velocity was measured at six different locations within the living area and balconies using a 3D
M

anemometer (WindMaster ultrasonic anemometer, Gill instrument), two air velocity transducers

(8475 series, TSI), and three 2D anemometers (WindSonic ultrasonic anemometer, Gill instrument).
D

Sensors where not re-calibrated in-situ as factory calibration was considered acceptable for this
TE

study. Specifications of each sensor are given in Table 1. Sampling rates of 1-5 Hz was used (Table

1), with collected data time averaged over 1-minute intervals for the detailed analysis. Sensors
EP

were installed 1.2m above the floor level representing the height of a seated human head. Two sets
C

of measurements for single-sided and cross ventilation were carried out for 24 hours for each

configuration during summer (13th and 14th January 2016). Measurements were recorded on both
AC

the balconies and in the living area. Doors to both the bedrooms and bathrooms were kept shut

during all measurements to minimise the impact of these on the ventilation in the main living area.

Figure 3 shows the positioning of all sensors for each configuration mode.

Page 5
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 3: case study plan and sampling location for cross ventilation (left) and single-sided ventilation (right)
M

Meteorological data from the Australian Government Bureau of Meteorology [21] from the closest
D

weather station to the case study building (Brisbane station) was used as reference weather
TE

conditions. The Brisbane weather station height is 8.13m and is located approximately two

kilometres from the case study building in an area with a similar urban setting to the case study.
EP

The obtained weather data included wind speed and direction, temperature and relative humidity

that were further used in the simulation validation and thermal comfort calculations.
C

Table 1: Sensors' specifications

AC

Instrument (manufacturer) NO. Parameters Accuracy and resolution Sampling rate

3D anemometer 1 U,V,W vectors Speed: <1.5% RMS @12 m/s 1Hz

(WindMaster ultrasonic Direction: 2° @12m/s

anemometer, model

Page 6
 ACCEPTED MANUSCRIPT

number 1590-PK-020, Gill

instruments)

2D anemometer 2 Wind speed and Speed: 2% @12m/s 4Hz

(WindSonic ultrasonic 2D direction or U Direction: 3° @12 m/s

PT
anemometer, Option 1, and V vectors

Serial Numbers 15170156

RI
and 15170157, Gill

instruments)

SC
2D anemometer 1 Wind speed and Speed: 2% @12m/s 4Hz

(WindSonic ultrasonic 2D direction or U Direction: 3° @12 m/s

U
anemometer, Option 4, and V vectors
AN
Serial Number13390047,

Gill instruments)
M

Velocity Transducer (8475- 2 Air velocity 3% of reading from 20° to 26° 5Hz

075-1, and 8475-150-1, C.

D

TSI) 1% of selected full scale range

TE

(2.5 m/s)
EP

2-2- Numerical method

2-2-1- CFD model and settings

C

CFD has been extensively applied for simulation of natural ventilation simulation in buildings [22-
AC

25]. The current study applied the 3D steady-state Reynold-Averaged Navier-Stokes (RANS) model.

RANS calculates the flow related parameters by solving time-averaged governing equations.

Despite some deficiencies [26, 27], RANS models are proven to be capable of simulating natural

ventilation reasonably well for both simple geometries [23, 28, 29] and buildings with detailed

façade elements such as balconies and double skin facades [30, 31]. Among the RANS models

Page 7
 ACCEPTED MANUSCRIPT

available, renormalisation group (RNG) k-ε performs better for ventilation simulations compared to

the others [24]. The RNG κ-ε turbulence model has been successfully used for simulation of both

indoor and outdoor airflows [28, 32-34] and has therefore been used in the present study. The RNG

κ-ε model is similar to the standard κ-ε model with a number of additional refinements that makes

PT
it more reliable for a different range of flows [35]. A comprehensive description of RNG κ-ε model

can be found in [35, 36]. Enhanced wall treatment was also implemented in this study to improve

RI
accuracy in the near-wall regions. ANSYS Fluent [37] combines enhanced wall function with the

SC
two-layer model at near-wall regions which results in a near-wall modelling approach that can

accurately calculate the flow in near-wall regions with relatively coarser meshes [35]. Wall

U
functions allow for coarser grids in the near-wall region, thus saving computational time [38], and
AN
have been used in the literatures for similar flows [39, 40].

The Archimedes number, Ar, (Equation 1) [41] was used to determine the relative dominancy of
M

wind and buoyancy forces in both the single-sided and cross ventilation cases:
D

∆
= (1)
TE

Where g is the gravitational acceleration (m/s2), H is the opening height (m), ΔT is the temperature

difference between inside and outside (K), T is the average temperature (K), and U is the average
EP

wind speed (m/s) at building height. Numbers less than 1 indicate dominancy of forced convection

over buoyancy induced natural convection, thus, buoyancy forces can be neglected with no
C

significant loss of accuracy [42].

AC

For this study, the Archimedes number for single-sided and cross ventilation are 0.018 and 0.015

respectively, indicating that the flow within the space is dominated by forced convection, and the

effects of buoyancy driven flow can be ignored. As such, only mass and momentum equations has

been solved and energy equation was not calculated.

Page 8
 ACCEPTED MANUSCRIPT

A CFD commercial code, ANSYS Fluent 17.0 [37] was employed to perform the simulations. The

SIMPLEC algorithm was adopted for pressure-velocity coupling, and the spatial discretization was

set to second-order upwind. Convergence was assumed to be achieved when all the residuals

reached the convergence criteria of 10-4.

PT
2-2-2- Numerical grids and computational domain

RI
A full-scale 3D model of the case study building was placed in a calculation domain. The domain size

was defined according to the case study building height (H) and has dimensions L × W × H= 1800 ×

SC
600 × 600 m3. Upstream and downstream lengths of the domain were 3H and 15H respectively,

lateral sides were 6H, and the domain height was equal to 6H (Figure 4). The domain dimensions

U
were defined based on the recommended values by the best practice guideline [43].
AN
Tong et al. [44] suggest three layers of obstructions in the CFD model to capture street canyon
M

effects, however, this requirement can be reduced for buildings at height while still capturing the

influence of surrounding obstructions [42]. Inclusion of surrounding obstructions, however, also

D

leads to more complicated flow patterns, and thus computational time. For this study, no
TE

surrounding building were included. This is considered an adequate simplification of the external

environment as there are no major obstructions with 250m on the southern side, and the primary
EP

focus of the study was the detailed comparison of the internal flow environment in relation to

different balcony parameters and ventilation modes. The authors recognise that the exclusion of
C

surrounding building will impact the airflow. However, this effect will be the same among all the
AC

case studies, and hence for a comparative study does not detract from the results.

An unstructured mesh with tetrahedral volume was created using ICEM CFD [45]. Two layers of 1.5

and 2.5 mesh density were applied to radius of 40m and 100m of the case study building

respectively. Grid refinements were applied to the openings and building surfaces. Opening meshes

were the most refined areas in the domain and consisted of grids with maximum sizes of 2e-3m.

Page 9
 ACCEPTED MANUSCRIPT

Maximum mesh size for the building surfaces ranged from 0.5m on surfaces distant from the

openings, 0.1m at surfaces adjacent to the case study unit, 4e-2m at the case study unit surfaces, 9e-

3m at the wall adjacent to the openings, to 6e-3m at the openings frames.

Grid sensitivity tests were performed by creating three sets of coarse, medium, and fine meshes

PT
with 8, 13, and 24 million elements respectively. Air velocity at key locations corresponding to

RI
experimental data points was used to assess the quality of the mesh. A 4% difference in results

between the coarse and medium mesh and a 1.6% difference between medium and fine mesh were

SC
obtained. Given this, the medium mesh with 13 million elements was considered to be able to

provide grid independent solutions and was used for the simulations.

2-2-3- Boundary conditions

U
AN
A power law equation (Equation 2) was used to calculate the wind boundary layer profile and the
M

acquired data was applied to the inlet boundary condition.

D

= ( ) (2)
TE

Where Vz (m/s) is wind speed at height z (m), Vref (m/s) is the reference velocity at the reference

height zref (m), and α is a component that is representative of terrain roughness. Considering the
EP

case study building is located in Brisbane CBD, α was set to 0.35 corresponding to city centre

terrains [46].
C
AC

Page 10
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 4: CFD domain size

Before the parametric study was performed the CFD model was validated against the experimental
M

data (explained in section 2-2-4). For the validation model, the inlet boundary condition was

defined according to the wind condition at the time of the experiment. Reference velocity values
D

were extracted from the reference weather station wind data associated with the time of the
TE

experiment. For the parametric study, however, the annual average wind speed (2.8 m/s) was used

as the reference velocity at the reference height of 8.13m (weather station’s height). Turbulence
EP

intensity was set to 5% representing a medium intensity and turbulent viscosity ratio was 10.
C

For all simulations, the outlet boundary condition was set to outflow, top and lateral boundaries
AC

were set to symmetry, and wall boundary conditions were applied to ground and building’s

surfaces as suggested by the best practice guideline [43]. In addition, all the wall boundary

conditions were no-slip wall with no additional surface roughness.

Page 11
 ACCEPTED MANUSCRIPT

2-2-4- CFD model validation

The results from CFD simulations were compared to the experimental data for both single-sided

and cross flow configurations. This was performed by considering the incident wind direction

perpendicular to the openings from the experimental data set. For this validation, internal data

PT
corresponding to an incident southerly wind were extracted, time averaged over 1-minute intervals

RI
and used to validate the computational results. The extracted external wind measurements (from

the reference weather station) where used as the inlet condition in the computational model. A

SC
number of 186 and 203 1-minute data values at each measurement point and reference weather

station were attained for single-sided and cross ventilation respectively. The obtained data were

U
averaged and measured values of air velocity inside the case study were compared to the
AN
simulation results of the same coordinates. Figure 5 shows the discrepancy between the simulation

results and experimental data for cross ventilation (Figure 5-A) and single-sided ventilation (Figure
M

5-B) as well as measurement uncertainties. As can be seen, CFD results slightly overestimate air

velocity for both cases (1-11%). The discrepancy between the CFD results and the experimental
D

data can be the resulted from steady-state assumption used in the simulations. Natural ventilation
TE

is unsteady in nature and usually involves wind fluctuations [19, 47]. The steady-state assumption

is likely to underpredict the fluctuation effect of wind [48], hence, the discrepancy between the
EP

simulation results and the experimental data is considered likely to be a result of the steady-state

assumption. These errors, however, are considered acceptable since significantly higher errors
C

have been reported in some similar studies [4, 16, 49].

AC

Page 12
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 5: comparison of measurement and simulation results for A) cross ventilation, and B) single-side

ventilation

U
AN
2-3- Tests configurations (case studies)

Balconies can be identified by two main features: depth and type. Therefore, to evaluate the effect
M

of a balcony on natural ventilation and indoor air flow, simulation cases were generated with

different balcony depths and types. Four different balcony depths expressed as a percentage of the
D

living area’s length (10%, 20%, 30% and 40%), as well as two balcony types (open balcony and
TE

semi-enclosed balcony), were defined. Figure 6 illustrates these variables. The defined balcony

variations were tested with both single-sided and cross ventilation configurations. In addition, to
EP

consider the effect balconies have under different incident winds, the case studies were tested

under four wind directions (0˚, 45˚, 90˚, and 180˚). Including the cases without a balcony, a total of
C

72 combinations of the aforementioned variables were formulated and tested. A summary of the
AC

configuration parameters is presented in Table 2. It needs to be noted that except for the varied

parameters, all the other parameters such as opening size and space length were kept constant.

Table 2: Configuration parameters

Variable Variations

Page 13
 ACCEPTED MANUSCRIPT

Balcony type Open Balcony (OB), Semi-enclosed Balcony (SB)

Ventilation mode Cross ventilation (CV), Single-sided Ventilation (SSV)

Balcony depth 0% (without balcony), 10%, 20%, 30%, 40%

Wind direction 0˚,45˚, 90˚, 180˚

PT
RI
U SC
d
AN
M

ɵ
D

Figure 6: Balcony types, open balcony (left) and semi-enclosed balcony (right)
TE

2-4- Thermal comfort model

EP

Comfort zone boundaries can be extended by elevating indoor air velocity [50, 51]. The

recommended model for prediction of thermal conditions for cases with the indoor air speed of
C

greater than 0.2 m/s is SET* [52]. ASHRAE standard-55 [52] defines SET* as the temperature of an
AC

environment at 50% relative humidity and average air speed of below 0.1 m/s, where air

temperature and radiant temperature are equal in which “the total heat loss from the skin of an

imaginary occupant with an activity level of 1.0 met and a clothing level of 0.6 clo is the same as

that from a person in the actual environment, with actual clothing and activity level”. In the current

study, nearly all the studied cases had represented air velocity of greater than 0.2 m/s, therefore,

Page 14
 ACCEPTED MANUSCRIPT

the SET* model was adopted for thermal comfort evaluation. The SET* model accounts for the

combined effect of temperature, humidity, air velocity, metabolic rate, and clothing insulation on

thermal comfort of the occupants [53]. Since changes in air velocity as a result of balcony provision

is of interest of this study, the remaining components were set to constant. Temperature and

PT
humidity values for January was extracted from the acquired meteorological data. The occupants

were assumed to be involved in a sedentary activity, metabolic rate, therefore, was set to 1.2 met.

RI
Light-weighted clothing corresponding to summer condition was also assumed, hence, 0.5 clo

SC
clothing value was adopted. The climatic conditions, obtained air velocity, metabolic rate, and

clothing insulation values were then used as inputs for SET* calculations using WinComf program

U
[54].
AN
3- Results and discussion

3-1- Results summary

M

Indoor air flow of the case studies was obtained using the validated CFD model. Average velocity in
D

the living area volume was extracted from the results and corresponding SET* values were
TE

calculated. Results are presented using average velocity and thermal comfort. The obtained velocity

results are summarised in Figure 7 and are categorised based on the investigated variables. The
EP

presented results for each variable (x-axis) is cumulative results from all the simulated cases where

the parameter of interest is looked at independently. For instance, the OB results are extracted from
C

all the cases with open balcony with different ventilation modes, depths, and wind angles.
AC

What stands out in Figure 7-A is that there is a significant difference between the average velocity

range when cross ventilation operates compared to that of the single-sided ventilation (about 7

times higher). To reveal the effects of the parameters on air velocity of each ventilation mode, cross

ventilation and single-sided ventilation results are plotted separately on Figure 7-B and Figure 7-C

respectively. As can be seen, both ventilation modes respond similarly to the change of variables in

Page 15
 ACCEPTED MANUSCRIPT

most cases: average velocity decreases as balcony depth increases, in terms of wind angle, the

highest velocity is achieved when the wind is normal to the openings whereas the lowest velocity is

associated with the wind parallel to the openings (90˚). An open balcony indicates a better

performance compared to the semi-enclosed balcony for both ventilation modes. However, for

PT
single-sided ventilation, the addition of an open balcony can result in an increase of air velocity

compared to the cases without balcony (0%), while in cross ventilation both balcony types result in

RI
a lower velocity than the cases without balcony. Therefore, there is a potential for improvement of

SC
single-sided ventilation through the addition of a balcony.

To look into the results in more detail, the average velocity results for all the simulated cases are

U
further presented in the following sections for single-sided and cross ventilation separately.
AN
M
D
TE
C EP
AC

Page 16
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 7: Results summary for both ventilation modes (A), cross ventilation (B), and single-sided

ventilation(C)

Page 17
 ACCEPTED MANUSCRIPT

3-1-1- Single-sided ventilation

The average velocity of the cases with single-sided ventilation are presented in Figure 8. The results

for different balcony depths are categorised based on the balcony types and the prevailing wind

directions. As can be seen in Figure 8, in overall, cases with semi-enclosed balcony present lower

PT
air velocities than the cases with open balcony for all the prevailing wind directions. Furthermore,

RI
the addition of semi-enclosed balconies have resulted in lower average velocities than the cases

without a balcony. In contrast, provision of open balconies have improved the average velocity in

SC
most cases. An increase in balcony depths have resulted in decrease of average velocity in the cases

with semi-enclosed balconies. This also applies to most of the cases with open balconies except for

U
the instances of 90˚ incident winds. In such cases, the average velocity increases with an increase in
AN
balcony depth of up to 30% and decreases for 40% depth. In terms of prevailing wind direction,

highest average velocity is achieved when the wind is perpendicular to the openings (0˚), followed
M

by 45˚, 180˚, and 90˚ incident winds respectively.

D

To summarise, the addition of a semi-enclosed balcony decreases the indoor average velocity,
TE

whereas, provision of an open balcony can improve the ventilation performance in most cases. The

open balcony can improve the average velocity significantly (up to 6 times) for the most
EP

unfavourable prevailing wind direction (90˚). The results also highlight the importance of building

orientation in ventilation performance of single-sided ventilation, where placing the openings

C

toward the prevailing wind direction captures at least twice air velocity as in other orientations.
AC

Page 18
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 8: Indoor average velocity for single-sided ventilation subject to various balcony type, depths and

prevailing wind direction.

Page 19
 ACCEPTED MANUSCRIPT

3-1-2- Cross ventilation

Average velocity results for cross ventilation are presented in Figure 9 based on different balcony

types, depths, and prevailing wind directions. Similar to single-sided ventilation, in cross ventilated

cases semi-enclosed balconies present a lower velocity average than open balconies. Both balcony

PT
types, however, result in a reduction of average velocity compared to the cases without balcony.

RI
The increase in balcony depth decreases the average velocity in most cases. The discrepancy

resulted by depth increase in cross ventilation, however, is up to 9.5% on average which is

SC
noticeably lower than that of the single sided ventilation (31% on average). The prevailing wind

direction also affects the average velocity in cross ventilation with the highest velocity

U
corresponding to the wind perpendicular to the openings (0˚), following by oblique (45˚) and
AN
parallel (90˚) wind directions.

It needs to be mentioned that despite the potential improvement of natural ventilation in single-
M

sided ventilation by the addition of an open balcony, cross ventilated cases still perform
D

significantly better (at least twice) than the improved single-sided ventilation cases.
TE
C EP
AC

Page 20
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP

Figure 9: Indoor average velocity for cross ventilation subject to various balcony type, depths and prevailing
AC

wind direction.

3-2- Sensitivity analyses

To identify the separate and combined impact of the balcony type, depth and the wind angle on

ventilation performance, a sensitivity analysis for each ventilation mode was conducted.

Page 21
 ACCEPTED MANUSCRIPT

To this end, a base case for each ventilation mode was selected and the other cases were compared

to the base case. The configurations with the lowest average velocity were chosen as the baseline

cases. Natural ventilation sensitivity was expressed as a percentage of increase in average velocity

as a result of altering the investigated variables comparative to the baseline case (Equation 3).

PT
Va-Vb
SA= V × 100 (3)
b

RI
Where SA is the sensitivity percentage, Va (m/s) is average velocity of configuration a, and Vb (m/s)

SC
is average velocity of the baseline case. Varying parameters for the analysis are balcony type (BT),

balcony depth (D), wind angle (W), and the combination of these independent variables (BT+D,

U
BT+W, D+W, and BT+D+W). The baseline cases for single-sided ventilation and cross ventilation
AN
were the case with 10% length semi-enclosed balcony under 90˚ wind direction (SB10-90), and the

case with 40% length semi-enclosed balcony under 90˚ wind direction (SB40-90) respectively.
M

Figure 10 shows the distribution of air velocity inside and around the baseline cases of single-sided

and cross ventilation configurations. Figure 10-A represents the air velocity distribution at the
D

unit’s height around the case study building for single-sided ventilation (left) and cross ventilation
TE

(right). As can be seen, the obstruction caused by the extended wall at the right side of the building

have induced airflow from right side to the left in cross ventilation case. It needs to be noted that
EP

Figure 10-C (single-sided baseline case) only represents the airflow inside the living area and

excludes the balcony. Due to a relatively higher air velocity inside the balcony, it was not possible to
C

capture air movement in both balcony and living area using the same scale for velocity.
AC

The sensitivity analysis was conducted for single-sided and cross ventilation separately and results

are discussed below.

Page 22
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 10: A) Velocity magnitude around the building for baseline cases of single-sided ventilation (left) and

cross ventilation (right), B) Velocity magnitude plan at 1.2m (top) and section A-A (bottom) for cross

ventilation baseline case, and C) Velocity magnitude plan at 1.2m (top) and section A-A (bottom) for single-

sided ventilation baseline case.

Page 23
 ACCEPTED MANUSCRIPT

3-2-1- Single-sided ventilation

Sensitivity percentage of average velocity for the investigated variables for single-sided ventilation

is presented in Figure 11. As can be seen, indoor average velocity is mainly affected by the incident

wind angle (W) followed by the balcony type (BT) and balcony depth (BD) respectively. Among the

PT
two varying parameters, air velocity is least sensitive to the combination of balcony type and

RI
balcony depth (BT+D) while reasonable improvements can be achieved by a combination of

balcony type with wind angle (BT+W). In addition, change in all the parameters simultaneously

SC
improves the ventilation performance but is not as effective as the change in balcony type and wind

angle. This highlights the high sensitivity of indoor air velocity to the approaching wind direction

U
and the minimal effect of balcony depth.
AN
Median
1800
M

1600
1400
1200
SA (%)

1000
D

800
600
TE

400
200
0
BT D w BT+D BT+W D+W BT+D+W
EP

Variables configurations
C

Figure 11: Sensitivity percentage of average air speed to different variables for single-sided ventilation
AC

3-2-2- Cross ventilation

It can be seen in Figure 12 that indoor air velocity is most sensitive to the change of incident wind

angle (W) followed closely by balcony depth (D) and is least sensitive to the balcony type (BT).

Additionally, varying two parameters together does not significantly change the SA compared to

one varying parameter and there is a small difference between SA of single and two parameters

Page 24
 ACCEPTED MANUSCRIPT

(6% between the best and the worst configurations). However, the most improvement can be

achieved by changing all the variables simultaneously.

Comparing the sensitivity analysis of cross ventilation and single-sided ventilation shows that

single-sided ventilation is much more sensitive to the change of the investigated parameters

PT
compared to cross ventilation. Altering different parameters in single-sided ventilation results in

RI
approximately 300% improvement on average while this number is about 50% for the cross

ventilation for the same variables. Besides, the mean SA discrepancy between different variable

SC
configurations for the cross ventilation is around 10% while in single-sided ventilation this number

is about 50 times higher (~ 500%). The suction effect caused by the pressure difference between

U
inlet and outlet of the cross ventilated cases is very dominant making this case less sensitive to the
AN
change of different variables compared to the single-sided ventilation. The average velocity in

single-sided ventilation is very low that even small configuration changes result in a notable change
M

of average velocity compared to the baseline case. It can also be seen that among the independent

parameters, both ventilation modes are most sensitive to the wind direction change. Although wind
D

direction cannot be controlled by building designers, buildings can be oriented in a way to take the
TE

most advantage of outside wind conditions.

C EP
AC

Page 25
 ACCEPTED MANUSCRIPT

Median
140
120
100
80
SA(%)

PT
60
40
20

RI
0
BT D w BT+D BT+W D+W BT+D+W
Variables configuration

SC
Figure 12: Sensitivity analyses of average air speed to different variables for cross ventilation configuration

U
It needs to be noted that the results presented in Figure 11 and Figure 12 are calculated using
AN
average velocity at the 1.2m plane. The same trends were also found using average velocity at 0.6m

and 1.8m planes as well as the living area’s volume.

M

3-3- Thermal comfort analyses

D

The results of average velocity in the living area of the case studies were used in the SET* index
TE

calculation. As could be expected, SET* values in the cases with cross ventilation were significantly

lower than that of the single-sided ventilation (3.4˚C on average). Differences in SET* values of
EP

various cases (ΔSET*) can be interpreted as the cooling effect on the human body. The potential

cooling effect of the investigated variables for the single-sided and cross ventilation configurations
C

are presented in Figure 13. It is evident from these results that the SET* values in the cases with
AC

single-sided ventilation respond more to the change of the variables compared to that of the cases

utilised with cross ventilation. In addition, regardless of ventilation mode, the approaching wind

angle has the most influence on potential cooling effect on the occupants compared to the balcony

depth and balcony type. It follows by balcony type in single-sided ventilated cases and balcony

Page 26
 ACCEPTED MANUSCRIPT

depth for the cross ventilated cases. These results are in accord with the results from sensitivity

analyses of the average velocity.

BT

PT
Variables

RI
W

SC
0 0.5 1 1.5 2 2.5 3 3.5
SET* (˚C)

Cross ventilation Single-sided ventilation

U
AN
Figure 13: investigated parameters potential cooling effect

In addition, ΔSET* of minimum and maximum values for single-sided and cross ventilation are 3.2˚C
M

and 1.1˚C respectively. This indicates that single-sided ventilation responds more to the change of
D

variables compared to the cross ventilation highlighting a higher potential for improvement.
TE

It needs to be noted that cross ventilation can provide adequate ventilation rate and thermal

conditions independent of changing variables, thus is more likely to create year-round comfort
EP

compared to single-sided ventilation which is heavily dependent on variables particularly wind

direction and building orientation.

C

4- Conclusion
AC

In-situ full-scale measurements of air velocity were conducted in a high-rise residential apartment.

The collected data was used to validate a CFD model from which a detailed investigation of the

separate and combined effect of the balcony type and depth, ventilation mode, and the wind angle

on indoor ventilation was performed. Various case studies were formulated based on two balcony

Page 27
 ACCEPTED MANUSCRIPT

types, four balcony depths, two ventilation modes and four wind angles. Average velocity and SET*

index were used as criteria and the following results were found:

• An open balcony results in a higher indoor velocity compared to the semi-enclosed balcony

PT
and natural ventilation performance of single-sided ventilation can be improved (up to

80%) by provision of open balcony.

RI
• The increase in balcony depth leads to decrease in air velocity.

• Among the tested incident wind directions, highest indoor air velocity corresponded to

SC
when the wind is normal to the openings, and is lowest when the prevailing wind is parallel

to the openings. This highlights the importance of orientating building toward the

U
prevailing wind direction for natural ventilation improvement.
AN
• Sensitivity analyses revealed that among wind angle, balcony depth, and balcony type, both

ventilation modes are most sensitive to the change of wind direction. It was also found that
M

the effect of altering the investigated parameters on natural ventilation performance is

D

much greater in single-sided ventilation (300% on average) compared to the cross

ventilation (50% on average). This emphasises on the importance of appropriate design in

TE

the case of single-sided ventilation.

• Among the two varying parameters, the most improvement was achieved by changing wind
EP

direction and balcony type in the case of single-sided ventilation. This highlights the
C

significant effect of wind direction which can be translated to the building orientation in
AC

building design. In cross ventilation, however, the most improvement was associated with

changing all the parameters together.

• The cooling effect on the human body (ΔSET*) shows changing the parameters in single-

sided ventilation has the maximum potential of 3.2 ˚C cooling effect improvement, while,

this number was found to be 1.1˚C for cross ventilation.

Page 28
 ACCEPTED MANUSCRIPT

Comparing single-sided and cross ventilation under the same circumstances shows a significantly

better natural ventilation performance in the case of cross ventilation. Analyses of different

parameters, however, reveals that single-sided ventilation is much more sensitive to the

parameters alteration compared to the cross ventilation. Therefore, considering lower performance

PT
of single-sided ventilation, additional care much be given to its design to assure the ventilation

effectiveness. Since implementation of cross ventilation is not always a possible solution, especially

RI
in dense urban areas, findings of this study provide solutions for improvement of single-sided

SC
ventilation design through appropriate choice of balconies. Having said that, the provided results

can also be used for the improvement of cross ventilation through design.

5- Limitations and future work

U
AN
The main aim of this study is to provide comparative results about the effect of different balcony

attributes on natural ventilation performance of high-rise residential units. Similar to any other
M

study, this study has some limitations that need to be addressed in future research. The current
D

study is carried out with an isothermal assumption and buoyancy-driven ventilation is neglected
TE

due to dominant effect of wind. In future studies, buoyancy-driven ventilation can be considered to

evaluate the effect of temperature gradient and buoyancy forces on ventilation performance of the
EP

balcony mounted residential units.

Acknowledgements
C

Computational and data visualisation resources used in this work were provided by the HPC and
AC

Research Support Group, Queensland University of Technology, Brisbane, Australia.

This research did not receive any specific grant from funding agencies in the public, commercial, or

not-for-profit sectors.

Page 29
 ACCEPTED MANUSCRIPT

Reference

[1] W. Liping, W.N. Hien, Applying natural ventilation for thermal comfort in residential buildings in
Singapore, Architectural Science Review 50(3) (2007) 224-233.
[2] M. Santamouris, F. Allard, Natural ventilation in buildings: a design handbook, Earthscan1998.
[3] R. Kennedy, L. Buys, E. Miller, Residents’ Experiences of Privacy and Comfort in Multi-Storey
Apartment Dwellings in Subtropical Brisbane, Sustainability 7(6) (2015) 7741-7761.

PT
[4] C. Gao, W. Lee, Evaluating the influence of openings configuration on natural ventilation
performance of residential units in Hong Kong, Building and Environment 46(4) (2011) 961-969.
[5] C. Mak, J. Niu, C. Lee, K. Chan, A numerical simulation of wing walls using computational fluid

RI
dynamics, Energy and Buildings 39(9) (2007) 995-1002.
[6] S. Omrani, B. Capra, V. Garcia-Hansen, R. Drogemuller, Investigation of the effect of balconies on
natural ventilation of dwellings in high-rise residential buildings in subtropical climate, Living and
Learning: Research for a Better Built Environment: 49th International Conference of the

SC
Architectural Science Association 2015, The Architectural Science Association, Melbourne,
Australia, 2015.
[7] C.-R. Chu, B.-F. Chiang, Wind-driven cross ventilation with internal obstacles, Energy and
Buildings 67(0) (2013) 201-209.

U
[8] A. Aflaki, N. Mahyuddin, Z. Al-Cheikh Mahmoud, M.R. Baharum, A review on natural ventilation
applications through building façade components and ventilation openings in tropical climates,
AN
Energy and Buildings 101 (2015) 153-162.
[9] R.J. Kennedy, L. Buys, Dimensions of liveability: a tool for sustainable cities, Proceedings of
SB10mad Sustainable Building Conference, 2010.
[10] L. Buys, J. Summerville, L. Bell, R. Kennedy, „Exploring the Social Impacts of High-Density
M

Living: Considering the Residents‟ Perspective.‟ In, Proceedings of Subtropical Cities 2008: From
fault-lines to sight-lines Subtropical urbanism in 20-20, 3–6 September 2008 (2008).
[11] J. Niu, Some significant environmental issues in high-rise residential building design in urban
D

areas, Energy and Buildings 36(12) (2004) 1259-1263.

[12] D. Song, Y.-J. Choi, Effect of building regulation on energy consumption in residential buildings
TE

in Korea, Renewable and Sustainable Energy Reviews 16(1) (2012) 1074-1081.

[13] E.A. Mohsen, D.J. Oldham, Traffic noise reduction due to the screening effect of balconies on a
building façade, Applied Acoustics 10(4) (1977) 243-257.
[14] I. Chand, P. Bhargava, N. Krishak, Effect of balconies on ventilation inducing aeromotive force
EP

on low-rise buildings, Building and Environment 33(6) (1998) 385-396.

[15] Z. Ai, C. Mak, J. Niu, Z. Li, Q. Zhou, The effect of balconies on ventilation performance of low-rise
buildings, Indoor and Built Environment (2011) 1420326X11409457.
[16] Z. Ai, C. Mak, J. Niu, Z. Li, Effect of balconies on thermal comfort in wind-induced, naturally
C

ventilated low-rise buildings, Building Services Engineering Research and Technology 32(3) (2011)
277-292.
AC

[17] E. Prianto, P. Depecker, Characteristic of airflow as the effect of balcony, opening design and
internal division on indoor velocity: A case study of traditional dwelling in urban living quarter in
tropical humid region, Energy and Buildings 34(4) (2002) 401-409.
[18] E. Prianto, P. Depecker, Optimization of architectural design elements in tropical humid region
with thermal comfort approach, Energy and Buildings 35(3) (2003) 273-280.
[19] L.J. Lo, A. Novoselac, Cross ventilation with small openings: Measurements in a multi-zone test
building, Building and Environment 57 (2012) 377-386.
[20] S. Omrani, V. Garcia-Hansen, R. Drogemuller, B. Capra, Thermal comfort evaluation of natural
ventilation mode: case study of a high-rise residential building, Fifty years later: Revisiting the role
of architectural science in design and practice: 50th International Conference of the Architectural

Page 30
 ACCEPTED MANUSCRIPT

Science Association 2016, Architectural Science Association and The University of Adelaide,
Adelaide, Australia, 2016.
[21] Australian Government Bureau of Meteorology, 2016. http://www.bom.gov.au/. (Accessed
January 2016).
[22] S. Omrani, R. Drogemuller, V. Garcia-Hansen, B. Capra, Natural ventilation heuristics in high-
rise residential buildings: evaluation and prediction, Across: Architectural Research through to
Practice: 48th International Conference of the Architectural Science Association (ANZAScA) 2014,

PT
The Architectural Science Association and Genova University Press, 2014, pp. 609-618.
[23] R. Ramponi, B. Blocken, CFD simulation of cross-ventilation for a generic isolated building:
impact of computational parameters, Building and Environment 53 (2012) 34-48.
[24] Q. Chen, Ventilation performance prediction for buildings: A method overview and recent

RI
applications, Building and Environment 44(4) (2009) 848-858.
[25] S. Omrani, V. Garcia-Hansen, B. Capra, R. Drogemuller, Natural ventilation in multi-storey
buildings: Design process and review of evaluation tools, Building and Environment 116 (2017)

SC
182-194.
[26] D. Lakehal, W. Rodi, Calculation of the flow past a surface-mounted cube with two-layer
turbulence models, Journal of Wind Engineering and Industrial Aerodynamics 67 (1997) 65-78.
[27] C.-H. Hu, M. Ohba, R. Yoshie, CFD modelling of unsteady cross ventilation flows using LES,

U
Journal of Wind Engineering and Industrial Aerodynamics 96(10) (2008) 1692-1706.
[28] G. Evola, V. Popov, Computational analysis of wind driven natural ventilation in buildings,
AN
Energy and Buildings 38(5) (2006) 491-501.
[29] J. Perén, T. Van Hooff, B. Leite, B. Blocken, CFD analysis of cross-ventilation of a generic isolated
building with asymmetric opening positions: impact of roof angle and opening location, Building
and Environment 85 (2015) 263-276.
M

[30] H. Montazeri, B. Blocken, CFD simulation of wind-induced pressure coefficients on buildings

with and without balconies: validation and sensitivity analysis, Building and Environment 60
(2013) 137-149.
D

[31] W. Pasut, M. De Carli, Evaluation of various CFD modelling strategies in predicting airflow and
temperature in a naturally ventilated double skin façade, Applied Thermal Engineering 37 (2012)
267-274.
TE

[32] M.Z.I. Bangalee, S.Y. Lin, J.J. Miau, Wind driven natural ventilation through multiple windows of
a building: A computational approach, Energy and Buildings 45(0) (2012) 317-325.
[33] Z.T. Ai, C.M. Mak, Determination of single-sided ventilation rates in multistory buildings:
Evaluation of methods, Energy and Buildings 69(0) (2014) 292-300.
EP

[34] R. Jin, J. Hang, S. Liu, J. Wei, Y. Liu, J. Xie, M. Sandberg, Numerical investigation of wind-driven
natural ventilation performance in a multi-storey hospital by coupling indoor and outdoor airflow,
Indoor and Built Environment 25(8) (2015) 1226-1247.
C

[35] A. Fluent, Release 15.0, Help System, Theory Guide, ANSYS, Inc.
[36] S.A. Orszag, V. Yakhot, W.S. Flannery, F. Boysan, D. Choudhury, J. Maruzewski, B. Patel,
AC

Renormalization group modeling and turbulence simulations, Near-wall turbulent flows (1993)
1031-1046.
[37] A. Fluent, Release 17.0, ANSYS, Inc., PA 15317, 2016.
[38] B. Blocken, T. Defraeye, D. Derome, J. Carmeliet, High-resolution CFD simulations for forced
convective heat transfer coefficients at the facade of a low-rise building, Building and Environment
44(12) (2009) 2396-2412.
[39] K.S. Nikas, N. Nikolopoulos, A. Nikolopoulos, Numerical study of a naturally cross-ventilated
building, Energy and Buildings 42(4) (2010) 422-434.
[40] K.A. Papakonstantinou, C.T. Kiranoudis, N.C. Markatos, Numerical simulation of air flow field in
single-sided ventilated buildings, Energy and Buildings 33(1) (2000) 41-48.

Page 31
 ACCEPTED MANUSCRIPT

[41] Z. Ai, C. Mak, D. Cui, On-site measurements of ventilation performance and indoor air quality in
naturally ventilated high-rise residential buildings in Hong Kong, Indoor and Built Environment
(2013) 1420326X13508566.
[42] A. Malkawi, G. Augenbroe, Advanced building simulation, Routledge2004.
[43] J. Franke, A. Hellsten, H. Schlünzen, B. Carissimo, Best Practice Guideline for the CFD Simulation
of Flows in the Urban Environment. COST Action 732: Quality Assurance and Improvement of
Microscale Meteorological Models (2007), Hamburg, Germany (2007).

PT
[44] Z. Tong, Y. Chen, A. Malkawi, Defining the Influence Region in neighborhood-scale CFD
simulations for natural ventilation design, Applied Energy 182 (2016) 625-633.
[45] ICEM CFD, ANSYS, Inc., PA 15317, 2016.
[46] A.G. Davenport, Rationale for determining design wind velocities, ASCE Journal of the

RI
Structural Division 86(5) (1960) 39-68.
[47] Y. Jiang, Q. Chen, Effect of fluctuating wind direction on cross natural ventilation in buildings
from large eddy simulation, Building and Environment 37(4) (2002) 379-386.

SC
[48] L.J. Lo, D. Banks, A. Novoselac, Combined wind tunnel and CFD analysis for indoor airflow
prediction of wind-driven cross ventilation, Building and Environment (2012).
[49] X. Liu, J. Niu, M. Perino, P. Heiselberg, Numerical simulation of inter-flat air cross-
contamination under the condition of single-sided natural ventilation, Journal of Building

U
Performance Simulation 1(2) (2008) 133-147.
[50] E.A. Arens, R. Gonzalez, L. Berglund, Thermal comfort under an extended range of
AN
environmental conditions, ASHRAE Transactions 92 (1986).
[51] D.R. Ernest, F.S. Bauman, E.A. Arens, The effects of external wind pressure distributions on
wind-induced air motion inside buildings, Journal of Wind Engineering and Industrial
Aerodynamics 44(1) (1992) 2539-2550.
M

[52] ASHRAE, Standard 55-2013, Thermal Environmental Conditions for Human Occupancy,
American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc., Atlanta, Ga
(2013).
D

[53] A.P. Gagge, A. Fobelets, L. Berglund, A standard predictive index of human response to the
thermal environment, ASHRAE Trans.;(United States) 92(CONF-8606125-) (1986).
[54] M. Fountain, C. Huizenga, WinComf: A Windows 3.1 Thermal Sensation Model—User’s Manual,
TE

Berkeley: Environmental Analytics (1996).

C EP
AC

Page 32
 ACCEPTED MANUSCRIPT

• The effect of balcony type, depth, ventilation mode, and wind direction on ventilation
performance was investigated.
• The addition of balcony to cross ventilation reduces the indoor air velocity.
• Provision of an open balcony to single-sided ventilation improves the ventilation performance.
• Ventilation performance of single-sided ventilation is mostly affected by the prevailing wind
direction.

PT
RI
U SC
AN
M
D
TE
C EP
AC