Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61

Contents lists available at ScienceDirect

Journal of Wind Engineering

and Industrial Aerodynamics
journal homepage: www.elsevier.com/locate/jweia

Air pollutant dispersion around high-rise buildings under different angles of MARK
wind incidence
⁎
Y. Yua,1, K.C.S. Kwokb,c, X.P. Liud,1, Y. Zhange,
a
School of Aerospace Engineering, Beijing Institute of Technology, China
b
Institute for Infrastructure Engineering, Western Sydney University, Australia
c
School of Civil and Environmental Engineering, Hong Kong University of Science and Technology,China
d
School of Civil Engineering, Hefei University of Technology, China
e
School of Clinical Medicine, Tsinghua University, China

A R T I C L E I N F O A BS T RAC T

Keywords: Gaseous pollutant originated from a building can disperse with ambient air flow and re-enter the same building
Air pollutant under different environmental conditions. In our previous studies, it was revealed that the near-wall dispersion
High-rise building characteristics in a windward emission are significantly different from that in leeward emission for a high-rise
Wind direction building due to the effect of the building on the wind flow. Since the atmospheric wind changes constantly in
Re-entry
both wind speed and wind direction, such that the wind rarely blows perpendicular to the front or the back of a
CFD
building, it is important to investigate what occurs under different angles of wind incidence. By means of both
physical wind tunnel measurement and simulations using Computational Fluid Dynamics (CFD), we firstly
studied the dispersion characteristics around a square-sectioned building. It was found that the building
influences on wind flow plays a significant role in the dispersion characteristics and an angle of wind incidence
of about 90 degrees is a transition angle for the pollutant dispersion pathway around the building. For angles of
wind incidence smaller than 90 degrees, air pollutant will migrate predominantly downward while for angles of
wind incidence greater than 90 degrees, the pollutant will migrate predominantly upward. Further tests
conducted on a more complicated crucifix-form building model with a re-entry along each building wing showed
a similar trend of pollutant dispersion at different angles of wind incidence, with the transition angle shifted to
approximately 75–80 degrees. The findings of this study show that building shape and the resultant wind-
structure interaction plays a significant role in the pollutant dispersion around a building, thus influencing the
air quality at different part of a building.

1. Introduction
In an urban environment, indoor air quality is strongly dependent
on urban air pollution. Indoor air quality in a building can be
compromised by the re-ingestion of contaminated exhaust air origi-
nated from the same building (Gao et al., 2009). Thus the character-
istics of gaseous pollutants dispersion around a building are especially
important for building design and indoor air evaluation.
Near-field dispersion of air pollutant not only affects general health
issues, it may also promote the spreading of communicable diseases
(Gao et al., 2009). It has been shown that several emerging and re-
emerging respiratory infections including tuberculosis (TB) (Nardell
et al., 1991) measles (Riley et al., 1978) and H1N1 influenza (Abd
Razak et al., 2013) can be spread by airborne route. Epidemiologic
analysis and simulation studies supported the possibility of wind effect
enhancing an airborne spread of the SARS virus in a large community
outbreak (Brook et al., 2004). The airborne transmission over long
distances can also be an important mode of infectious disease
transmission (Nardell et al., 1991). These studies have increased the
awareness of the scientific and engineering communities to micro-scale
pollutant dispersion in the built environment.
Evidently, air ventilation of urban areas by wind flow is an
important process in urban pollutant drainage to maintain clean air.
In particular, outdoor air quality can be improved by wind flow because
wind dilutes and removes pollutants. Urban wind flow is strongly
related to urban morphology as a combination of building density,
mutual arrangement of buildings and their individual shape and
dimensions. Hence, wind-structure interaction, which controls air

⁎
Correspondence to: School of Clinical Medicine, Tsinghua University, Beijing 100084, China.
E-mail address: yuzhang2014@tsinghua.edu.cn (Y. Zhang).
1
Those authors have equal contributions to this paper.

http://dx.doi.org/10.1016/j.jweia.2017.04.006
Received 3 September 2016; Received in revised form 13 March 2017; Accepted 11 April 2017
0167-6105/ © 2017 Elsevier Ltd. All rights reserved.
Y. Yu et al.
(a) Crucifix-form building model in boundary layer wind tunnel
(b) Simplified square-sectioned building
model in a numerical wind tunnel
(H:B:D=1:0.98:1.04)
Fig. 1. Physical and numerical models.
pollutant dispersion around buildings, needs to be properly considered
and the results incorporated in the decision making process for urban
development programs.
During the past three decades, a considerable amount of research
has been conducted regarding air pollutant dispersion in urban
environments by wind action, including field measurements, wind
tunnel tests and numerical simulations (Depaul and Sheih, 1985, 1986;
Nakamura and Oke, 1988; Meroney et al., 1996; Baik et al., 2000; Lu
et al., 2012). Researchers have found that, from a macroscopic point of
view, the dispersion of air pollution is generally determined by
meteorological factors such as ambient wind speed, wind direction,
and atmospheric stability. In addition to these natural factors, urban
design, including the arrangement of building arrays and street
canyons, also plays an important role in air pollutant dispersion. Yet,
in the neighborhood within the local turbulent boundary layer close to
the buildings where wind turbulence, vortex formation, exhaust gas
emission and natural convection all play a role, the air pollutant
dispersion process is far more complex. These neighborhood regions
can be crucial as unexpected air flow may result in long residual time
for air pollutant close to the residents which further promotes disease
spreading. However, which factor amongst natural convection, wind
turbulence, and wind-structure interaction plays the most significant
role in such areas remains a controversial issue (Niu and Tung, 2008;
Yip et al., 2007; Zhou and Jiang, 2004; Li et al., 2005).
In previous work (Liu et al., 2010, 2011; Zhang et al., 2015), we
52
Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61
(c) Crucifix-form building model
in a numerical wind tunnel
(scale 1:30)
have investigated the wind effect on the pollutant dispersion around a
typical Hong Kong high-rise residential (HRR) building with a crucifix
platform and a re-entry along each wing. A series of tests using a scaled
model was studied in an atmospheric boundary layer wind tunnel
(Dimoudi and Nikolopoulou, 2003; Tominaga et al., 2008). Numerical
simulations were also performed. The results obtained from both
experimental measurement and simulations highlighted the differences
in air pollutant dispersion pathway for a windward or leeward emission
source, and the dependence on the location of the emission source in
relation to the stagnation point. In the following discussion, the
releasing point is below the stagnation. In the case of a windward
emission, pollutant migrates predominantly downward due to
downwash effect and spreads horizontally after reaching the ground.
For a leeward emission, air pollutant migrates predominantly upward
within the re-entry due to the pressure field and flow recirculation
around the building before discharging downstream. Obviously the air
pollutant dispersion process is sensitive to the wind direction; hence it
is important to investigate how pollutant dispersion pathway responds
to a change in the angle of wind incidence from windward (0°) to
leeward (180 degrees). Furthermore, the effect of building shape on air
pollution dispersion, particularly near-building dispersion remains
relatively under-explored.
The objectives of this research were to investigate air pollutant
dispersion around high-rise buildings of different shapes under differ-
ent angles of wind incidence. The emphasis of this research was to
Y. Yu et al. Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61

investigate the mechanisms behind the transition of the pollutant

dispersion pathway from predominantly downward to predominantly
upward. The results will advance our understanding of the effect of
wind-structure interaction on air pollutant dispersion around high-rise
buildings. This advancement of knowledge will influence future high-
rise building design and urban planning to enhance air pollution
dispersion and minimize disease transmission around high-rise build-
ings.

2. Method

2.1. Wind tunnel model study

A 1:30 scale model was constructed to represent a 10 storey high-

(a) rise building 30 m in height in prototype with a crucifix shape typical of
residential buildings in Hong Kong, as shown in Fig. 1(a). Based on the
Hong Kong building code, each building wing must have a “re-entry” to
comply with building regulation that requires all rooms to be fitted with
a window/opening. The experimental study was carried out in a
boundary layer wind tunnel at the CLP Power Wind/Wave Tunnel
Facility (WWTF) at Hong Kong University of Science and Technology
(HKUST). The test section is 41 m in length, 4 m in height and 5 m in
width. Experiments were conducted to study the pollutant dispersion
around this complex building structure. The blockage ratio in the wind
tunnel is controlled at approximately 5% to minimize the requirements
for blockage correction. For wind tunnel modeling of flow and plumes
dispersion, similarity requirements between scale model and prototype
have been summarized by Meroney and others (Meroney, 2004).
Amongst these similarity conditions, Reynolds number (Re) indepen-
dence is one of the important requirements. When the Reynolds
(b)
number is sufficiently large, both flow and scalar dispersion can be
considered effectively independent of Re. During the experiment, the
Reynolds number calculated based on the building dimension and
approach wind velocity measured at building height was approximately
2.2×105, which is sufficiently high to ensure that the measurement
results are effectively independent of Reynolds number.
The tracer gas used was propane in air with a 99,000 ppm propane
concentration. The tracer gas was released through a flow-meter at a
constant flow rate of 58.5 ml/s. The emission exit opening was
enlarged to ensure the exit velocity of the source gas was below
0.4 m/s, which was about one order of magnitude lower than the wind
speed. Taking into consideration the density difference between the
source gas and ambient air, the local buoyancy Froude number is
approximately 2.69. The tracer gas was emitted at the same position
within the re-entry at 3rd floor, independent of the angle of wind
(c)
incidence.
A Cobra probe, with an accuracy of ± 0.3 m/s and ± 1° pitch and
k (m2/s2)
0.0 0.2 0.4 0.6 0.8 1.0
yaw up to about 30% turbulence intensity, was used to determine the
1.4
Wind profile in Exp
wind profile. The detection of tracer gas concentrations was achieved
1.2 Wind speed in CFD with a fast-response flame ionization detector (FID) (Cambustion
Turbulence kinetic energy in CFD
Model HFR400), which was carefully calibrated using synthetic air
1.0 Turbulence dissipation rate in CFD
and certified calibration gases of different concentrations (from
51.2 ppm to 99,000 ppm). The calibration was repeated to ensure that
Z (m)

0.8

0.6
the calibration error is generally within 10% when it was used to
measure the standard gas with the lowest concentration, which is
0.4 51.2 ppm. It was found that there were some daily variations in the FID
0.2 calibration. Hence the calibration was performed before and after each
series of experiment to check the stability of the equipment. A pre-
0.0
0 1 2 3 4 tested analysis was performed to estimate the minimum sampling time
U (m/s), (m2/s2) from each specific test. The minimum averaging time to reach a
(d) converged (i.e. representative) standard deviation was between 60 s
Fig. 2. CFD settings. and 90 s at selected key measurement points. Therefore, at each
measurement position, sample air was collected over a period of
120 s at a data-acquisition rate of 150 Hz to obtain stable estimates
of mean concentration. The signal was then passed through a low-pass
filter which effectively eliminated the alias errors (Liu et al., 2010,

53
Y. Yu et al. Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61

103 100

Normalized Concentration Kc
Normalized Concentration Kc
102 80

101 60

100 40
180 degree Experiment
10-1 180 degree k-e model 20 180 degree-coarsesca mesh
180 degree k-w sst model
180 degree LES 180 degree-medium mesh
180 degree-fine mesh
10-2
0 2 * 4 6 8 10
0
0 2 * 4 6 8 10
Building height: Floor Building height: Floor
(a) Simulations using different CFD models (b) Comparision of results from the
three grids

Fig. 3. CFD models and grid independence.

2011). Measurement of tracer gas concentration at each floor within
the re-entry area was carried out after the wind flow has stabilized.
Details of the experimental work can be found in the previous reports
(Liu et al., 2010, 2011).
2.2. Computational Fluid Dynamic (CFD) study
Two building models were simulated using CFD. One model
replicated the outer profile of the crucifix-form building to generate a
square-sectioned building model, which is shown in Fig. 1(b). This
method of simplification is consistent with the simulation of building
groups in urban area (Dimoudi and Nikolopoulou, 2003; Tominaga
et al., 2008), which provides a considerable saving in computational
time. The second model is a replica of the crucifix-form model which
maintains the size and configuration identical to the physical model
tested in the wind tunnel, as shown in Fig. 1(c). ANSYS Fluent CFD
Software Package (version 14.0) was used for this study.
2.2.1. Computational domain and grid
The distances from the buildings to the side, to the inlet and to the
top of the domain are at least 5 H, with H is the height of the building.
The distance from the building to the outlet is 15 H. Therefore, the U (z ) =
requirements by Tominaga are satisfied (Tominaga et al., 2008).
The computational grid was non-uniform and fully structured. A
maximum stretching ratio of 1.2 was applied to reduce the grid density
in areas far away from the region of interest. The grid system is
presented in Fig. 2 for both building models. In order to study the grid
independence, the baseline computational mesh is coarsened and
refined with twice about a factors 2. This produces three grids of
successive refinement with 1,671,780 cells, 3,411,958 cells and
5,986,120 cells. The grid resolution resulted from a grid-sensitivity
analysis that will be outlined later.
2.2.2. Boundary conditions
As shown in Fig. 2(d), the measured approaching velocity in the
wind tunnel test can be described by a logarithmic law (Eq. (1)), where
u*/ABL is determined based on the reference velocity at the building
height (3.27 m/s at 1 m) (Liu et al., 2010; Peren et al., 2015a) κ is the
von Karman constant (0.42) and z the height coordinate. Based on the
curve-fitting, the aerodynamic roughness length z 0=0. 66mm (equiva-
lent to 0.02 m in prototype, at a scale of 1:30). The turbulent kinetic
energy k was calculated from the mean wind speed and the measure
turbulence intensity using Eq. (2), where Iu is the stream wise
turbulence intensity and α is a parameter assumed to be 1.0 here.
The turbulence dissipation rate ε was given by Eq. (3) and the specific
dissipation rate ω by Eq. (4), where Cμ is an empirical constant taken to
be equal to 0.09.
u*/ ABL z + z0
ln ( )
κ z (1)
k (z ) = α (Iu (z ) U (z ))2 (2)
u*3/ ABL
ε (z ) =
κ (z + z 0 ) (3)

54
Y. Yu et al.
(a) 3-D view of the model rotation
θ
(b) Top view of the model rotation
The wind direction is specified, only the building model is rotated to generate an
angle of wind incidence (θ)
Fig. 4. Angle of wind incidence.
ε (z )
ω (z ) =
Cμ k (z )
For the ground surface, the standard wall functions, proposed by
Launder and Spalding (Launder BE, 1974) with roughness modifica-
tion by Cebeci and Bradshaw (Cebeci, 1977), were used. The values of
the roughness parameters, i.e. the sand-grain roughness height k s (m)
and the roughness constant CS , were determined using their consis-
tency relationship with the aerodynamic roughness length z 0 derived by
Blocken et al. Blocken et al. (2007)(Eq. (5)) for FLUENT and CFX.
9. 793z 0
ks =
CS
The value of k s is restricted in FLUENT; it should be smaller than
yP , which is the distance between the center point of the wall-adjacent
cell and the wall. Therefore, for the ground surface surrounding the
modelled buildings, k s is equal to 0.93 mm and consequently CS should
be 7 to satisfy Eq. (5) for z 0=0. 66mm . A user-defined function is used
to set the value of the constant CS outside the standard allowable
interval of [0;1] in FLUENT 14.0. The building surfaces were set to
have zero roughness height (k s=0m ), as recommended by (Peren et al.,
2015b).
Zero static pressure is imposed at the outlet plane and slip wall
conditions, i.e. zero normal velocity and zero normal gradients of all
variables, at the top and lateral sides of the domain.
A pollution source containing 10% propane with a 58.5 ml/s
volume flow rate was numerically emitted from within the re-entry at
3rd floor, which is located under the stagnation point, as marked by a
circle in Fig. 1(b) and (c). Mass flow rate boundary was used at the
pollution emission window.
The air pollutant concentrations at different locations were scaled
CU H2
based on a non-dimensional concentration KC = QH where C is the
computed mean concentration, UH is the mean wind speed at building
model height H , and Q is the volumetric flow rate of the tracer gas (Liu
et al., 2010).
Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61
2.2.3. Solver settings
The simulations were performed with the commercial CFD code
Fluent 14.0. The 3D steady RANS equation were solved using Standard
k-ε model. The SIMPLE algorithm was used for pressure-velocity
coupling, pressure interpolation was second order and QUICK dis-
cretization schemes were used for convection terms and second-order
center discretization schemes for viscous terms of all the governing
equations, including momentum, k and ε, species transportation
equations. Convergence was assumed to be obtained when all the
scaled residuals leveled off and reached a minimum of 10−6 for x, y and
z momentum, k, ε, and species.
2.2.4. Computational grid resolution and turbulence model choice
In order to select an appropriate turbulence model, the simulation
results of pollutant distribution using three widely accepted turbulence
models: unsteady Large Eddy Simulation (LES), k−ω− sst, and k−ε
model, were compared. The comparison with experimental results
(Fig. 3a) showed that the differences between these models in the
prediction of pollutant dispersion were not significant, which is
consistent with the results of a recently published study by Yuan
et al. The emission concentration can be more accurately modelled by a
LES model, but the computational cost of the LES model is several
times higher than that of the RANS model (Tominaga and
Stathopoulos, 2011). In this study, we focused on the influence of
wind direction on the pathway of air pollutant, which are mainly macro
behaviour determined by the mean flow field, balancing the computa-
tional cost and accuracy, thus k-ε model was selected to simulate the
air pollutant transportation (Patankar, 1980).
Three grid sizes of approximately 1,671,780 cells, 3,411,958 cells
and 5,986,120 cells were used to study gird sensitivity through the
(4) simulations of the leeward emission case for high-rise building. The
results for the three grid sizes are shown in Fig. 3(b). For the grid-
convergence study, we adopted the concept of the grid-convergence
index (GCI) by Roache (Roache, 1994, 1997). The calculated GCI is
shown in Fig. 3(c), which confirmed that the medium sized grid is
sufficient for the numerical simulations reported in this paper. In order
to balance the simulation accuracy and computation time, we chose the
medium level of 3,411,958 cells for all the following simulations, in
which the smallest cell dimension is 5 mm×5 mm×5 mm in the x-, y-,
and z- directions at closed to the edges of the buildings, and
2.5 mm×2.5 mm×2.5 mm at close to the pollutant releasing location.
(5)
2.3. Rotation of building model in the CFD simulation
A wind direction θ=0° is referenced to Y positive direction, as
shown in Fig. 4(a). A change in wind direction was achieved by rotating
the building model about the Z direction in an anticlockwise direction
to create a corresponding angle of wind incidence, as shown in
Fig. 4(b). The pollutant emission point was fixed and rotated with
the building model.
In order to avoid the interpolation of CFD results, which is
necessary when a dynamic mesh is used to represent the building
model rotation, a new set of grid was regenerated each time the
building model was rotated. To reduce uncertainty introduced by every
new grid for different wind directions, we separate the domain into two
zones, one is the central region including building, which is 1 H away
from the building model side surface and 4 H from building model top
surface. Another region is the rest of the domain. For every new set of
grid, the central region was rotated and the mesh distribution in this
central region including building remained unchanged, only the
meshes outside this central region were changed, but not too much.”
Thus, in total there were 26 sets of CFD grid for 2 models; 13 wind
directions for each model by rotating every 15 degrees from 0 to 180
degrees.
55
Y. Yu et al.
Fig. 5. Air pollutant dispersion around the square-sectioned model at different angles of wind incidence.
3. Result and discussion
3.1. Pollutant concentration around square-sectioned model
Fig. 5 shows the pollutant concentration around the building at
different angles of wind incidence. At θ=0° where the emission is at the
windward face, pollutant dispersed predominantly downward due to
downwash, as shown in Fig. 5(a). As the angle of wind incidence was
progressively shifted away from 0°, for example, at θ=30° as shown in
Fig. 5(b), the flow regime around the building was controlled by an
increasing lateral flow component. At θ=90° where the emission is at
the side face, the pollutant became entrained within the recirculation
56
Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61
zone in the near wall region, which is dominated by the separation-
reattachment mechanism created by flow separation along the wind-
ward corner. Characteristically, the air flow, and hence pollutant
concentration, show a distinctive form of flow toward upstream,
particularly close to the building surface.
Evidently, θ=90° signifies a transition angle of wind incidence
beyond which the pollutant starts to migrate upwards. At angles of
wind incidence greater than θ=90°, the emission is at the leeward face
where the air flow is dominated by the highly turbulent flow within the
wake region. Furthermore, the wake flow associated with air flows over
and around the building is also characterized by a large flow recircula-
tion with significant vertical flow components that creates an updraft
Y. Yu et al. Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61

(a) θ=0° Air pollutant migrates downwards due to downwash

(b) θ=90° Air pollutant migrates horizontally upstream due to recirculating flow
underneath the separated shear layer

(c) θ=180° Air pollutant migrates upwards due to large recirculation bubble
Fig. 6. Flow pattern and directions of air pollutant dispersion around the square-sectioned model at different angles of wind incidence.

close to the leeward face of the building. This upward air flow along the
leeward face results in the distribution features of building surface
pollutant concentration, as shown in Fig. 5(d) and (e) for θ=120° and
θ=150°, respectively. At θ=180°, the air flow, and hence the pollutant
concentration, display a nearly vertically upward pattern, as shown in
Fig. 5(f).
To further illustrate the occurrence of the transition angle of wind
incidence at around θ=90°, Fig. 6 displays the streamlines around the
pollutant emission point under different angles of wind incidence. At
θ=0° (windward emission), since the emission point is below the
stagnation point, the dominant downwash drives the pollutant down-
ward toward the ground, as shown in Fig. 6(a).
At θ=90° (side wall emission), the air flow in the near wall region is
predominantly toward upstream in the horizontal direction, as shown
in Fig. 6(b). This migration of pollution upstream against the incident
wind is driven primarily by the recirculating flow underneath the
separated shear layer originating from the windward edge which
reattaches intermittently along the building side face (Saathoff and
Melbourne, 1989).
At θ=180° (leeward emission), in the near wake adjacent to the
leeward face, the air flow is predominantly upward, as shown in
Fig. 6(c). This upward flow in the near wake is associated with the
flow and pressure fields in the wake of the building characterized by a
large recirculating bubble stretches far downstream (Tsang et al.,
2012).
Comparing the streamlines and air pollutant dispersion pathways,
it is clearly evident that the change in pollutant dispersion pathway
corresponds to the air flow pattern driven by wind-structure interaction
at different angles of wind incidence. The near-wall pollutant disper-
sion route changes from “vertical downward”, to “horizontal backward”
and then to “vertical upward” when the incident wind angles changes
from 0° to 90° and then to 180°.

57
Y. Yu et al.
Fig. 7. Air pollutant dispersion around the crucifix-form building with re-entries at different angles of wind incidence.
3.2. Pollutant concentration around crucifix-form building model
Fig. 7 shows the pollutant concentration around the crucifix-form
building model at different angles of wind incidence, which can be
compared with that for the square-sectioned building model. Despite
differences associated with the different building shapes, the pollutant
concentration distribution is predominantly governed by the air flow
pattern produced by wind-structure interaction. At θ=0°, pollutant
migrated predominantly along the re-entry and downward due to
downwash, before spilling outward at close to ground level and
dispersed, as shown in Fig. 7(a). At θ=30°, initially pollutant was
entrapped and migrated downward within the re-entry before escaping
58
Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61
the re-entry close to ground level and dispersed, as shown in Fig. 7(b).
At θ=90°, the entrapped pollutant migrated upward prior to exiting the
re-entry, as shown in Fig. 7(c). Pollutant dispersion was dominated by
wake flow at angles of wind incidence greater than 90°, such as at
θ=120°, 150° and 180° shown in Fig. 7(d), (e) and (f) respectively.
Evidently, not only was the pollutant driven upward, it remained
trapped within the re-entry until it exited at the roof and dispersed. It is
noteworthy that the re-entry of the crucifix-form building model acts as
a plenum where initial mixing and dilution of the emitted pollutant
take place. However, the re-entry also acts as a well-protected conduit
which traps emitted pollutant and restricts its interaction with ambient
air, thus hinders the dilution process. Hence the entrapped pollutant
Y. Yu et al.
Fig. 8. Air pollutant dispersion around the crucifix-form building with re-entries at a transient angle of wind incidence.
Fig. 9. Flow pattern within reentry at the transient angle of wind incidence.
remains high in concentration within the re-entry. Since the re-entry is
connected to existing areas within the building so that building
occupants may encounter pollutant within the re-entry, this presents
a grave health hazard if the pollutant contains a high concentration of
harmful substances.
It is likewise notable that although the flow pattern and the
pollutant concentration distribution of both the square-sectioned and
crucifix-form building models underwent a transition from predomi-
nantly downward to predominantly upward, the angle of wind inci-
dence at which the transition took place and the characteristics of the
transition differ between the two building models. In order to further
reveal the transition process when the angles of wind incidence is
between θ=75° and θ=90°, additional simulations were conducted at
θ=5° interval for further analysis. These differences are highlighted in
Fig. 8 which shows the pollutant concentration distribution at θ=75°
and 80°. A close examination of pollutant concentration in the vicinity
of the emission point suggests that the pollutant remained suspended
within the re-entry and the concentration value fluctuated slowly. The
flow pattern at θ=75° shown in Fig. 9, particularly the streamlines close
to the emission point, clearly shows a clear downdraft at θ=75° but
updraft at θ=80°, which means between 75° and 80° is the transient
angle of wind incidence to change pollutant concentration value
distribution trend inside re-entry.
Fig. 10 shows the computed and measured normalized tracer gas
concentration along the building height within the re-entry under
59
Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61
θ= 75°
θ= 80°
different angles of wind incidence. As shown in Fig. 10(a), for angles of
wind incidence within 0° to 60°, the concentration distribution within
the re-entry indicates a similar trend of dispersion where pollutant
predominantly migrates downwards. Experimental measurement, ta-
ken in the case with a windward emission, confirmed the simulation
results with a reasonable agreement. When the angle of wind incidence
is within 105° to 180°, the pollutant predominantly migrates upwards,
which is again supported by the experimental measurement of the
event with a leeward emission (θ=180°), as shown in Fig. 10(b). In the
transition range, i.e. angle of wind incidence is within 75° to 90°, the
maximum air pollutant concentration occurred at the emission point
(3rd floor), hence air pollutant concentration firstly increases and then
decreases as the floor number increases from 1st floor to 10th floor,
which indicates the suspension of the pollutant around the emission
point. Experiment conducted with a side wall emission (90 degrees)
also recorded a high concentration of air pollutant within the re-entry,
particularly around the emission point at 3rd floor, as presented in
Fig. 10(c). Overall, despite the variability of tracer gas concentration
measurements taken under experimental conditions, and the assump-
tions and simplifications made during numerical simulations, the
experimental results generally validated the simulation work and the
flow characteristic and pollutant concentration distribution at different
angles of wind incidence, including the transition angle of wind
incidence.
The results presented in this paper clearly demonstrated that the
Y. Yu et al. Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61

103 Exp-00 degree 102

00 degree
Normalized Concentration Kc

Normalized Concentration Kc
00 degree 30 degree
15 degree 101 60 degree
102 30 degree
45 degree 100
60 degree
101 10-1

10-2
100
10-3
-1
10
10-4

10-2
0 2 * 4 6 8 10
10-5
0 2 4 6 8 * 10
Building height: Floor Building height: Floor
(a) θ = 0° to 60° (a) θ = 0° to 60°

103 103
90 degree

Normalized Concentration Kc
Normalized Concentration Kc

120 degree
150 degree
102 102 180 degree

101 101

105 degree
100
100 120 degree
135 degree
150 degree 10-1
10-1 165 degree
180 degree
Kc-Exp-180degree 10-2 *
10-2
0 2 * 4 6 8 10
0 2 4 6 8
Building height: Floor
10

Building height: Floor

(b) θ = 90 to 180
(b) θ = 105° to 180°
Fig. 11. Normalized concentration distributions within the re-entry at different angles
103 of wind incidence for pollutant emission (* show pollutant emission point).
Normalized Concentration Kc

and a re-entry along each building wing. Despite the noticeable

102
differences in the building shape, the pollutant dispersion pathway
around the two buildings was found to show a similar tendency.
101 Generally, when the angle of wind incidence positions the pollutant
emission point at a windward location, the pollutant will migrate
downward due to downwash. When the angle of wind incidence
100
Exp-90degree positions the pollutant emission point at a leeward location, the
75 degree-revised pollutant will migrate upward due to the highly turbulent flow and
10-1 80 degree-revised flow recirculation within the wake region in the lee of the building.
85 degree-revised
90 degree-revised However, the local air flow and pollutant dispersion characteristics are
highly dependent on the local building features, for example, within
10-2
0 2
* 4 6 8 10 and in close proximity to the re-entry of the crucifix-form building
Building height: Floor model. The near-wall pollutant dispersion route changes from “vertical
(c) θ = 75° to 90° downward”, “horizontal backward” to “vertical upward” for the square-
sectioned building when the angle of wind incidence changes from 0° to
Fig. 10. Normalized concentration distributions within the re-entry at different angles 180°. However, for the crucifix-form building model, the results show
of wind incidence for 3rd pollutant emission (* show pollutant emission point).
competing downdraft and updraft which give rise to an oscillating flow
and a corresponding fluctuation in pollutant concentration value.
angle of wind incidence and the building shape play an important role
Furthermore, the transition angle at which the pollutant dispersion
in the pollutant dispersion process and thus influence the air quality at
pathway changes from predominantly downward to predominantly
different part of a building.
upward also differs slightly, from around 90° for the square-sectioned
building model to around 75° for the crucifix-form building model.
4. Conclusion It should be noted that the results presented above are dependent
on the actual location of the emission source on the building face. They
CFD simulations were performed to study the effect of angle of wind are indicative of the pollution dispersion pathways for different angles
incidence on air pollutant dispersion around a simplified square- of wind incidence for an emission location at the 3rd floor of a 10 storey
sectioned model and a similar height prototype building model of a building, which is below the stagnation point. To validate the effect of
typical Hong Kong high-rise residential building with a crucifix-form the emission location, we further tested a case in which the emission

60
Y. Yu et al.
location is at the 9th floor, which is above the stagnation point and very
close to the roof. The results show that for a wide range of angle of wind
incidence, the pollutant migrates upwards within the re-entry, as
shown in Fig. 11.
Acknowledgments
This research was supported by a special fund of the State Key Joint
Laboratory of Environment Simulation and Pollution Control
(2015)−15K09ESPCT, Tsinghua University, China, and the Research
Grant Scheme (RGS) 2014, Institute for Infrastructure Engineering,
Western Sydney University, Australia.
References
Abd Razak, A., Hagishima, A., Ikegaya, N., Tanimoto, J., 2013. Analysis of airflow over
building arrays for assessment of urban wind environment. Build. Environ. 59,
56–65.
Baik, J.J., Park, R.S., Chun, H.Y., Kim, J.J., 2000. A laboratory model of urban street-
canyon flows. J. Appl. Meteorol. 39, 1592–1600.
Blocken, B., Stathopoulos, T., Carmeliet, J., 2007. CFD simulation of the atmospheric
boundary layer: wall function problems. Atmos. Environ. 41, 238–252.
Brook, R.D., Franklin, B., Cascio, W., Hong, Y., Howard, G., Lipsett, M., Luepker, R.,
Mittleman, M., Samet, J., Smith, S.C., Jr., Tager, I., Expert Panel on, P., Prevention
Science of the American Heart, A., 2004. Air pollution and cardiovascular disease: a
statement for healthcare professionals from the Expert Panel on Population and
Prevention Science of the American Heart Association. Circulation 109, 2655–2671.
Cebeci, T., B.P., 1977. Momentum Transfer in Boundary Layers. Hemisphere Publishing
Corporation, New York.
Depaul, F.T., Sheih, C.M., 1985. A tracer study Of dispersion Inan urban street canyon.
Atmos. Environ. 19, 555–559.
Depaul, F.T., Sheih, C.M., 1986. Measurements of wind velocities in a street canyon.
Atmos. Environ. 20, 455–459.
Dimoudi, A., Nikolopoulou, M., 2003. Vegetation in the urban environment:
microclimatic analysis and benefits. Energ. Build. 35, 69–76.
Gao, N.P., Niu, J.L., Perino, M., Heiselberg, P., 2009. The airborne transmission of
infection between flats in high-rise residential buildings: particle simulation. Build.
Environ. 44, 402–410.
Launder BE, S.D., 1974. The numerical computation of turbulent flows. Comput. Meth
Appl. Mech. Eng. 3, 21.
Li, Y., Duan, S., Yu, I.T.S., Wong, T.W., 2005. Multi-zone modeling of probable SARS
virus transmission by airflow between flats in Block E. Amoy Gard., Indoor Air 15,
96–111.
Liu, X.P., Niu, J.L., Kwok, K.C.S., Wang, J.H., Li, B.Z., 2010. Investigation of indoor air
pollutant dispersion and cross-contamination around a typical high-rise residential
building: wind tunnel tests. Build. Environ. 45, 1769–1778.
Liu, X.P., Niu, J.L., Kwok, K.C.S., Wang, J.H., Li, B.Z., 2011. Local characteristics of
61
Journal of Wind Engineering & Industrial Aerodynamics 167 (2017) 51–61
cross-unit contamination around high-rise building due to wind effect: mean
concentration and infection risk assessment. J. Hazard Mater. 192, 160–167.
Lu, C., Deng, Q.H., Liu, W.W., Huang, B.L., Shi, L.Z., 2012. Characteristics of ventilation
coefficient and its impact on urban air pollution. J. Cent. South Univ. T 19, 615–622.
Meroney, R.N., 2004. Wind tunnel and numerical simulation of pollution dispersion: a
hybrid approach. Invited Lecture (in: Croucher Advanced Study Insitute on Wind
Tunnel Modeling). Hong Kong University of Science and Technology.
Meroney, R.N., Pavageau, M., Rafailidis, S., Schatzmann, M., 1996. Study of line source
characteristics for 2-D physical modelling of pollutant dispersion in street canyons.
J. Wind Eng. Ind. Aerod 62, 37–56.
Nakamura, Y., Oke, T.R., 1988. Wind, temperature and stability conditions in an East
West Oriented Urban Canyon. Atmos. Environ. 22, 2691–2700.
Nardell, E.A., Keegan, J., Cheney, S.A., Etkind, S.C., 1991. Airborne infection -
theoretical limits of protection achievable by building ventilation. Am. Rev. Respir.
Dis. 144, 302–306.
Niu, J., Tung, T.C., 2008. On-site quantification of re-entry ratio of ventilation exhausts
in multi-family residential buildings and implications. Indoor Air 18, 12–26.
Patankar, S.V., 1980. Numerical Heat Transfer and Fluid Flow. Hemisphere Publishing
Corporation, US.
Peren, J.I., van Hooff, T., Leite, B.C.C., Blocken, B., 2015a. Impact of eaves on cross-
ventilation of a generic isolated leeward sawtooth roof building: windward eaves,
leeward eaves and eaves inclination. Build. Environ. 92, 578–590.
Peren, J.I., van Hooff, T., Leite, B.C.C., Blocken, B., 2015b. CFD analysis of cross-
ventilation of a generic isolated building with asymmetric opening positions: impact
of roof angle and opening location. Build. Environ. 85, 263–276.
Riley, R.L., Riley, E.C., Murphy, G., 1978. Airborne spread of measles in a suburban
elementary-school. Am. Rev. Respir. Dis. 117, (255-255).
Roache, P.J., 1994. Perspective - a method for uniform reporting of grid refinement.
Stud., J. Fluid Eng.-T Asme 116, 405–413.
Roache, P.J., 1997. Quantification of uncertainty in computational fluid dynamics. Annu
Rev. Fluid Mech. 29, 123–160.
Saathoff, P.J., Melbourne, W.H., 1989. The generation of peak pressures in separated
reattaching flows. J. Wind Eng. Ind. Aerod 32, 121–134.
Tominaga, Y., Stathopoulos, T., 2011. CFD modeling of pollution dispersion in a street
canyon: comparison between LES and RANS. J. Wind Eng. Ind. Aerod 99, 340–348.
Tominaga, Y., Mochida, A., Yoshie, R., Kataoka, H., Nozu, T., Yoshikawa, M., Shirasawa,
T., 2008. AIJ guidelines for practical applications of CFD to pedestrian wind
environment around buildings. J. Wind Eng. Ind. Aerod 96, 1749–1761.
Tsang, C.W., Kwok, K.C.S., Hitchcock, P.A., 2012. Wind tunnel study of pedestrian level
wind environment around tall buildings: effects of building dimensions, separation
and podium. Build. Environ. 49, 167–181.
Yip, C., Chang, W.L., Yeung, K.H., Yu, I.T., 2007. Possible meteorological influence on
the severe acute respiratory syndrome (SARS) community outbreak at Amoy
Gardens, Hong Kong. J. Environ. Health 70, 39–46.
Zhang, Y., Kwok, K.C., Liu, X.P., Niu, J.L., 2015. Characteristics of air pollutant
dispersion around a high-rise building. Environ. Pollut. 204, 280–288.
Zhou, Z.X., Jiang, C.Q., 2004. [Effect of environment and occupational hygiene factors of
hospital infection on SARS outbreak], Zhonghua lao dong wei sheng zhi ye bing za
zhi = Zhonghua laodong weisheng zhiyebing zazhi = Chinese journal of industrial
hygiene and occupational diseases, 22 261-263.