Content uploaded by Ming Fu
Author content
All content in this area was uploaded by Ming Fu on Jul 18, 2014
Content may be subject to copyright.
http://ibe.sagepub.com/
Indoor and Built Environment
http://ibe.sagepub.com/content/early/2013/09/23/1420326X13504122
The online version of this article can be found at:
DOI: 10.1177/1420326X13504122
published online 24 September 2013Indoor and Built Environment
Z. Y. Han, W. G. Weng, Q. Y. Huang, M. Fu, J. Yang and N. Luo
limbs pendulum and body motion
Aerodynamic characteristics of human movement behaviours in full-scale environment: Comparison of
Published by:
http://www.sagepublications.com
On behalf of:
International Society of the Built Environment
can be found at:Indoor and Built EnvironmentAdditional services and information for
http://ibe.sagepub.com/cgi/alertsEmail Alerts:
http://ibe.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
What is This?
- Sep 24, 2013OnlineFirst Version of Record >>
at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from at UNIV CALIFORNIA BERKELEY LIB on October 10, 2013ibe.sagepub.comDownloaded from
I
I
ndoor
ndoor
and
and B
uilt
uilt
Environment
Original Article
Aerodynamic characteristics
of human movement behaviours
in full-scale environment:
Comparison of limbs pendulum
and body motion
Z. Y. Han, W. G. Weng, Q. Y. Huang, M. Fu, J. Yang and N. Luo
Abstract
The aerodynamic effects of human movement can significantly influence the airflow motion and con-
taminants transmission in enclosed environments such as in aircraft cabins, cinema and conference
room, and so it is necessary and important to study the characteristics of these aerodynamic effects.
This work focuses on the aerodynamic characteristics of human movement behaviours including limbs
pendulum and body motion. A thermal manikin is used in the corresponding environments to simulate
these behaviours. The step frequencies of 20, 30, 40, 50 and 60 double steps per minute for limbs
pendulum and the moving speeds of 0.5, 0.75, 1.0, 1.25 and 1.5 m/s for body motion are investigated
in the experiments. In each case, the velocity distribution around the human body is measured by using
hot-wire anemometers. The experimental results show that the characteristics of the velocity distribu-
tion induced by limbs pendulum depend on pendulum frequency and spatial location, and for body
motion, it depends on moving speed, moving distance and spatial location. The analysis results show
that body motion has a significant influence on the flow field, and its aerodynamic effects are greater
than those due to limbs pendulum. When a human moves, more detailed profile of the human body
leads to more complicated flow field in the nearby area.
Keywords
Aerodynamic effect, Thermal manikin, Flow disturbance, Human movement, Limbs pendulum
Accepted: 14 August 2013
Introduction
For many years, respiratory infectious diseases such as
influenza and severe acute respiratory syndrome are
threatening the life of the humans around the world.
1
Every year, respiratory infectious diseases infect almost
one-quarter of the global population and lead to 4 mil-
lion deaths due to respiratory infections and 1.5 million
deaths due to tuberculosis.
2
A pandemic of avian flu
among humans can cost the global economy $800 bil-
lion a year.
3
Many epidemiology reports have indicated
that the respiratory infectious diseases may be spread
by airborne transmission within enclosed environment
such as isolation rooms and airliner cabins.
4–11
The air-
flow from front to back of the cabin (longitudinal) is
minimal due to the ventilation system in airplane,
which means in-flight transmission of a disease contam-
inant should be confined within two rows of an infected
passenger.
11–14
So, the infection of the passengers
seated far from the index patient might be more likely
because of the movements of the walking crew members
or passengers along the airline cabin. The aerodynamic
effects of these movements might result in the
Department of Engineering Physics, Institute of Public Safety
Research, Tsinghua University, Beijing, China
Corresponding author:
W. G. Weng, Department of Engineering Physics, Institute of
Public Safety Research, Tsinghua University, Beijing, PR China.
Email: wgweng@tsinghua.edu.cn
Indoor and Built Environment
0(0) 1–15
!The Author(s) 2013
Reprints and permissions:
sagepub.co.uk/
journalsPermissions.nav
DOI: 10.1177/1420326X13504122
ibe.sagepub.com
longitudinal airflow and enhance the dispersion of
contaminant.
15
Recently, studies on the aerodynamic effects of the
human movement have received more and more atten-
tion.
16
Many significant works have been made on this
issue including numerical simulation and experiments.
For numerical simulation, the computational fluid
dynamics (CFD) method has been widely used for the
quantitative calculation of the effects of human move-
ment in enclosed environment.
15–20
Although plenty of
studies were made on this issue, significant differences
could be found among the conclusions of these works.
Edge et al.
16
studied the unsteady aerodynamic wake of
a walking person by using a time-accurate CFD simu-
lation. The results indicated an apparent unsteady
bluff-body wake behind the torso and a region of
unsteady vortex shedding behind the legs. Shih et al.
17
studied the effect of human movement on room air dis-
tribution in an isolated room and concluded that the
removal of contaminants was not obviously affected by
the moving speed. Mazumdar et al.
18
investigated the
effect of human movement on contaminant concentra-
tion distributions in a single inpatient ward. The results
indicated that the average concentration change in the
breathing levels in the ward was generally small. But,
different from Shih et al.
17
and Mazumdar et al.,
18
Wang and Chow
19
studied the influence of human
walking on dispersion and deposition of expiratory
droplets in an isolation room. The results showed that
the human walking disturbed the local velocity field,
and the increases of walking speed could effectively
reduce the overall number of suspended droplets.
Similar conclusion was also given by Mazumdar
et al.,
15
in which the aerodynamic effects of human
movements were simulated in an airplane cabin. The
results showed that the movement of a crew member
or a passenger could carry a contaminant in its wake to
as many rows as the crew member or passenger
passed.
15
All the human bodies in these studies men-
tioned above were simple objects. Realistic walking
human model was used by Choi and Edwards
20
in
room compartments where a contaminated and a
clean room were connected by a vestibule. The contam-
inant transport owing to human movement and door
motions and vent-system activity was investigated, and
the results showed that faster walking speed resulted in
less mass transport from the contaminated room into
the clean room.
For experimental studies, Matsumoto and Ohba
21
studied the effects of moving object by using a movable
heated object in a full-scale room model in experiment.
Poussou et al.
22
studied the flow field by using a one-
tenth scale water-based model. The flow field and
contaminant transport were measured using the par-
ticle image velocimetry and planar laser-induced
fluorescence techniques, respectively. Bjorn and
Nielsen
23
studied the influence of the physical move-
ment and exhalation of human on the concentration
distributions of contaminant in full-scale test rooms
with life-sized breathing thermal manikins. The results
of these previous works showed that the movement of
individuals in enclosed environments could significantly
influence the contaminant transmission and personal
exposure to contaminants. However, for the experi-
ments in the small water tank model, changes in the
physical scale and working fluid further complicate
the interpretation of equivalent effects in the full
scale.
24
Until now, few full-scale experiments were per-
formed because experiments in full-scale environment
are highly dependent on the environmental airflow,
which is quite difficult to be controlled. For the experi-
ments in small full-scale model, the human objects were
regular-shaped objects rather than manikin with
human profile. In fact, the complicated profile of
the human may also influence the airflow around the
human.
25
So, experiments using manikin with human
profile are necessary for the investigation of airflow dif-
fusion induced by human movement. In addition,
human walking behaviours such as arm and leg pendu-
lum may also affect the near-by flow field in the differ-
ent ways, which were not investigated in the previous
works. Therefore, accurate measurements of the flow
field induced by human walking in full-scale environ-
ment are strongly needed for aerodynamic analysis.
In this work, experimental investigation was con-
ducted in full-scale chamber environment to study the
aerodynamic effects of human movement behaviours
including limbs pendulum and body motion on the
transmission of contaminants and diseases.
Experimental setup and method
In general, human movement behaviour involves the
pendulum of both limbs and body motion. In this
work, the characteristics of limbs pendulum and body
motion were separately investigated because the two
behaviours are difficult to be executed simultaneously
using a manikin in experiments.
Limbs pendulum
This experiment was conducted in a full-scale chamber
(6 m 6m2.6 m). The thermal manikin used in this
experiment was a thermal manikin ‘Newton’ (manufac-
tured by Measurement Technology Northwest, USA).
It was made of carbon fibre-epoxy body form with ther-
mally conductive reinforcement. This manikin was able
to simulate the physiological behaviours of real human
such as heat generation, sweating, breathing and walk-
ing. The profile of this manikin was well designed to
2Indoor and Built Environment 0(0)
accurately represent the shape of a real person, as
shown in Figure 1. The height of this manikin was
1.68 m, the width was 0.58 m and the length was
0.29 m. For the limbs of the manikin, the length of
the arms was 0.66 m (from the middle of the shoulder
joint to the end of the finger), and the length of the legs
was 0.88 m (from the middle of the hip to the soles of
the feet).
To simulate the walking behaviour of limbs pendu-
lum, a standard Manikin Walking Motion Stand of this
thermal manikin was used. When installed on the
Stand, the manikin was hung on and its feet were
0.28 m above the ground. The arms and legs were con-
nected with the Stand by fixing the hands and feet to
the traction devices of the Stand. The joint of shoulder
and hip of the manikin were also able to rotate flexibly.
By using this Stand, the walking motion of limbs pen-
dulum was achieved by periodically pulling the arms
and legs, as illustrated in Figure 1. The step frequency
was controlled by the control system of the Stand, and
the accuracy was 1 double step per minute (dspm).
When the walking behaviour was performed by the
Stand, the rotation angle of the arms was 60, and
the pendulum distance of the hands in the horizontal
direction was 0.66 m. For the legs, the rotation angle
was 50, and the pendulum distance of the feet in the
horizontal direction was 0.74 m. During the experi-
ment, the manikin and the Stand were set in the
middle of the chamber. The chamber was big enough
to reduce the limitation effects of the walls on the flow
field. When the limbs swung, the body of the manikin
did not move backward or forward, and the location of
the Stand did not change.
For measuring the velocity of the airflow induced by
the manikin movement, 24 one-dimensional hot-wire
anemometers were used (Kanomax, system 6242 with
model 1550, 1504 and velocity probe 0963-00 A200,
Japan), as shown in Figure 1. There were 12 anemom-
eters around the right arm of the manikin: six located
on the right side (sensors 1–6) and six in front (sensors
7–12), with the height corresponding to the upper arm,
forearm and wrist, respectively. Another 12 anemom-
eters were set around the left leg: six in front (sensors
13–18) and six on the left side (sensors 19–24), with the
height corresponding to the thigh, calf and ankle. The
distance between the anemometers and the limbs was
small but enough for avoiding any kind of impacts. The
location of the anemometers is given in Table 1. The
origin of the coordinate was located at the left side of
the human on the ground. The coordinate system met
the right-hand rule. The Xdirection was from the left
side to the right side of the manikin. The Ydirection
was from the back to the front of the manikin, and the
Zdirection was from the ground to the top. The loca-
tion of the top vertex on the head of the manikin was
X¼0.33 m, Y¼0.06 m and Z¼1.96 m. The location of
Figure 1. Experimental setup of limbs pendulum: (a) the thermal manikin and the Manikin Walking Motion Stand and
(b) schematic diagram of the experimental setup and the location of the anemometers.
Han et al. 3
the anemometers and the coordinate system can also be
seen in Figure 1(b). The anemometer 2, 4, 6 and 19, 21,
23 were closer to the manikin than the anemometer 1, 3,
5 and 20, 22, 24, respectively. For measuring the vel-
ocity ranging from 0.1 to 4.99 m/s, the velocity reso-
lution of the anemometer was 0.01 m/s and the
sampling frequency was 10 Hz. The measured data
was velocity magnitude in the form of scalar quantity.
One temperature and humidity detector (Vaisala
HMP60, Finland) was also used to monitor the tem-
perature and relative humidity in the chamber. Around
20C, the accuracy of the temperature and humidity
detector was 0.6C and 3% RH, and the response
time was 1 s. There was no ventilation or any other
moving objects in this chamber, so only the movement
of the manikin could cause airflow motion in this
environment.
Body motion
This experiment was conducted in a full-scale cabin
(24 m 2.3 m 2.6 m). The movement of the manikin
was achieved by using a double-track orbit
(10 m 0.56 m 0.09 m) and a trolley (0.6 m 0.6 m
0.165 m) located in the middle of the cabin. The ther-
mal manikin was fixed on the trolley set on the orbit. The
moving speed of the trolley was controlled by using a
voltage regulator. The speed of the trolley ranged from 0
to 1.5 m/s, and the acceleration and deceleration time of
this instrument was 0.3 s. Figure 2(a) is a photograph of
the thermal manikin and the orbit in the cabin, and
Figure 2(b) is a schematic diagram of the experimental
setup of body motion. The cabin was long enough to
reduce the limitation effects of the wall at the two ends
of the flow field.
Thirty hot-wire anemometers were used for measur-
ing the velocity of the airflow induced by body motion.
For every 1.5 m along the orbit, six anemometers were
placed as one group on the right side of the orbit, and
the first group was placed 3 m away from the beginning
of the orbit. For every six anemometers, the height cor-
responded to the chest, hip and knee, respectively. The
location of the anemometers is given in Table 2. The
origin of the coordinate system was located at the left
side of the orbit starting point on the ground. The
coordinate system met the right-hand rule. The Xdir-
ection was from the left side of the manikin to the right.
The Ydirection was from the back to the front of the
manikin, and the Zdirection was from the ground to
the top. The location of the top vertex on the head of
the manikin before the movement was X¼1.15 m,
Y¼0.5 m and Z¼1.96 m. The anemometers with odd
number were closer to the manikin than those with even
number. The location of the anemometers and the
coordinate system can also be seen in Figure 2(b). In
Xdirection, the distance between the anemometers and
the manikin was small but enough to avoid any kind of
impact. There were also five temperature and humidity
detectors, which monitor the temperature and relative
humidity in the corridor, close to each group of
anemometers. The location of the temperature and
humidity detectors is given in Table 3. There was no
ventilation or any other moving objects in this cabin, so
only the movement of the manikin could cause airflow
motion in this environment.
Experimental cases
For the experiment of limbs pendulum, the step fre-
quency was set as 20 dspm, 30 dspm, 40 dspm,
Table 1. Location of the anemometers in the experiment of limbs pendulum.
No. X(m) Y(m) Z(m)
Corresponding
location Area No. X(m) Y(m) Z(m)
Corresponding
location Area
1 0.78 0.025 1.50 Beside arm Upper arm 13 0.25 0.54 0.90 In front of leg Thigh
2 0.7 0.025 1.50 Beside arm Upper arm 14 0.17 0.54 0.90 In front of leg Thigh
3 0.78 0.025 1.30 Beside arm Forearm 15 0.25 0.54 0.70 In front of leg Calf
4 0.7 0.025 1.30 Beside arm Forearm 16 0.17 0.54 0.70 In front of leg Calf
5 0.78 0.025 1.10 Beside arm Wrist 17 0.25 0.54 0.50 In front of leg Ankle
6 0.7 0.025 1.10 Beside arm Wrist 18 0.17 0.54 0.50 In front of leg Ankle
7 0.63 0.42 1.50 In front of arm Upper arm 19 0.12 0.09 0.90 Beside Leg Thigh
8 0.55 0.42 1.50 In front of arm Upper arm 20 0.04 0.09 0.90 Beside Leg Thigh
9 0.63 0.42 1.30 In front of arm Forearm 21 0.12 0.09 0.70 Beside Leg Calf
10 0.55 0.42 1.30 In front of arm Forearm 22 0.04 0.09 0.70 Beside Leg Calf
11 0.63 0.42 1.10 In front of arm Wrist 23 0.12 0.09 0.50 Beside Leg Ankle
12 0.55 0.42 1.10 In front of arm Wrist 24 0.04 0.09 0.50 Beside Leg Ankle
4Indoor and Built Environment 0(0)
50 dspm and 60 dspm, respectively. Twenty double
steps per minute were used for simulating stroll, 60
dspm was for fast walking and 30 dspm, 40 dspm and
50 dspm were for simulating normal walking at differ-
ent speeds. For the experiment of body motion, the
moving speed was set as 0.5 m/s, 0.75 m/s, 1.0 m/s,
1.25 m/s and 1.5 m/s respectively. About 0.5 m/s was
used for simulating stroll, 1.5 m/s was for fast walking
and 0.75 m/s, 1.0 m/s and 1.25 m/s were for simulating
normal walking at different speeds. The cases investi-
gated in the experiments are shown in Table 4. Since the
step length of the manikin was 0.74 m during limbs
pendulum, the step frequency of 20–60 dspm of
cases 1–5 was corresponding to the walking speed of
Figure 2. Experimental setup of body motion: (a) the thermal manikin and the orbit in the cabin and (b) schematic diagram
of the experimental setup and the location of the anemometers.
Table 2. Location of the anemometers in the experiment of body motion.
No. X(m) Y(m) Z(m) Group
Corresponding
area No. X(m) Y(m) Z(m) Group
Corresponding
area
1 1.51 3.0 1.50 1 Chest 16 1.59 6.0 1.15 3 Hip
2 1.59 3.0 1.50 1 Chest 17 1.51 6.0 0.75 3 Knee
3 1.51 3.0 1.15 1 Hip 18 1.59 6.0 0.75 3 Knee
4 1.59 3.0 1.15 1 Hip 19 1.51 7.5 1.50 4 Chest
5 1.51 3.0 0.75 1 Knee 20 1.59 7.5 1.50 4 Chest
6 1.59 3.0 0.75 1 Knee 21 1.51 7.5 1.15 4 Hip
7 1.51 4.5 1.50 2 Chest 22 1.59 7.5 1.15 4 Hip
8 1.59 4.5 1.50 2 Chest 23 1.51 7.5 0.75 4 Knee
9 1.51 4.5 1.15 2 Hip 24 1.59 7.5 0.75 4 Knee
10 1.59 4.5 1.15 2 Hip 25 1.51 9.0 1.50 5 Chest
11 1.51 4.5 0.75 2 Knee 26 1.59 9.0 1.50 5 Chest
12 1.59 4.5 0.75 2 Knee 27 1.51 9.0 1.15 5 Hip
13 1.51 6.0 1.50 3 Chest 28 1.59 9.0 1.15 5 Hip
14 1.59 6.0 1.50 3 Chest 29 1.51 9.0 0.75 5 Knee
15 1.51 6.0 1.15 3 Hip 30 1.59 9.0 0.75 5 Knee
Han et al. 5
0.5–1.5 m/s of cases 6–10. In cases 6–10, the total
moving distance was 9 m, and the total moving time
for the manikin was 18 s, 12 s, 9 s, 7.2 s and 6 s,
respectively.
For each case, three measurements were operated in
which the anemometers were settled to face the X,Y,Z
direction, respectively. In these measurements, the hot-
wire probe was well-installed to keep the three directions
of the probe (normal to the wire in the prongs-probe
plane, along the wire and normal to the prongs-probe
plane) exactly following the direction of the three axes of
the coordinate system. Every measurement was repeated
three times, and the instantaneous velocity was mea-
sured. In cases 1–5, the measurements lasted 60 s. For
cases 6–10, the measurement time was 25 s, 20 s, 15 s, 15 s
and 15 s, respectively. Every measurement would not
begin unless the flow field in the experimental environ-
ment became static state and the airflow was undetect-
able (the velocities measured by the anemometers were
all zero). For avoiding the influences of the heat released
by the human on the flow field, the manikin was not
heated, and there was no heat source in the experimental
environment. The effects of the heat plume flow around
the human body were not considered. Only the move-
ment of the manikin could cause airflow motion in the
experimental environment and change the flow field. To
test and verify the stability of the temperature distribu-
tion and the relative humidity distribution in the experi-
mental environment, the temperature and relative
humidity in the full-scale environment were also
recorded during each measurement.
Calculation of velocity magnitude
In three-dimensional flow field, the results measured by
the hot-wire anemometers in this work are not the cor-
responding velocity components of the airflow but a
combination of them. Since the effective velocity mag-
nitude measured by anemometer depends on the
three velocity components of the flow field and the
direction of the anemometer setup, the corresponding
relationship between the effective velocity magnitude
VEmeasured by the hot-wire anemometer and the
three velocity components can be estimated by
equation (1):
26,27
V2
E¼V2
Uþk2V2
Vþh2V2
Wð1Þ
where VU,VVand VWare the velocity components in
different directions, m/s, Uis the direction normal to
the wire in the prongs-probe plane, Vis the direction
along the wire, Wis the direction normal to the prongs-
probe plane, kis the yaw factor associated with the
tangential velocity along the axis and his the pitch
factor associated with the velocity component normal
to the prongs-sensor plane.
For the anemometers used in this work, the mea-
sured direction was normal to the prongs-sensor
plane. So, in the calibration process before the experi-
ment, the anemometers were re-calibrated in a one-
dimensional flow field normal to the prongs-probe
plane, which means V2
e¼h2V2
m. So, equation (1)
becomes equation (2), as:
V2
m¼1
h2V2
Uþk2V2
Vþh2V2
W
ð2Þ
where Vmis the velocity magnitude measured by the
anemometers used in this work. By yaw calibration,
the value of kwas obtained as 0.12, and a nominal
value of 1.02 was used for h.
28
In each measurement,
the hot-wire probes were well-installed to keep U,V
and Wdirection exactly follow the directions of the
three axes of the coordinate system, respectively.
Assumed that the airflow motion induced by the
human movement has a good similarity in the repeated
measurement, the instantaneous velocity magnitude
VtðÞ of the airflow induced by human movement at
time tcan be obtained by using equation (3):
V2ðtÞ¼ V2
xðtÞþV2
yðtÞþV2
zðtÞ
0:5
¼h2
1þk2þh2
V2
m,xðtÞþV2
m,yðtÞþV2
m,zðtÞ
0:5
¼0:7116 V2
m,xðtÞþV2
m,yðtÞþV2
m,zðtÞ
0:5
ð3Þ
Table 3. Location of the temperature and humidity detec-
tors in the experiment of body motion.
No. X(m) Y(m) Z(m) Group
Corresponding
area
1 1.61 3.0 m 1.15 m 1 Hip
2 1.61 4.5 m 1.15 m 2 Hip
3 1.61 6.0 m 1.15 m 3 Hip
4 1.61 7.5 m 1.15 m 4 Hip
5 1.61 9.0 m 1.15 m 5 Hip
Table 4. The cases investigated in the experiment.
Case
no. Behaviour
Step
frequency
(dspm)
Case
no. Behaviour
Moving
speed
(m/s)
1 Limbs pendulum 20 6 Body motion 0.5
2 Limbs pendulum 30 7 Body motion 0.75
3 Limbs pendulum 40 8 Body motion 1.0
4 Limbs pendulum 50 9 Body motion 1.25
5 Limbs pendulum 60 10 Body motion 1.5
6Indoor and Built Environment 0(0)
where Vm,xtðÞ,Vm,ytðÞand Vm,ztðÞare the instantaneous
velocity magnitudes measured in X,Yand Zdirection
at time t, respectively, m/s. VxtðÞ,VytðÞand VztðÞare the
instantaneous velocity components of the airflow in X,
Yand Zdirection at time t, respectively, m/s. By using
equation (3), the velocity magnitude of the airflow
induced by the human movement in the experiments
can be obtained.
Results
Limbs pendulum
In this experiment, the temperature and relative humid-
ity around the manikin did not change with the limbs
pendulum of the manikin. During the repeated meas-
urements, the variation range of temperature was
0.2C and for relative humidity 0.4%, less than
the instrumental error of the detectors. This stability
was because the temperature and relative humidity dis-
tribution were uniform and in a stable state before each
measurement.
By observing the velocity distribution measured in
repeated measurements, similarity and repeatability
were established. So, the average of the results mea-
sured in the repeated measurement was used as
Vm,xtðÞ,Vm,ytðÞand Vm,ztðÞ. Then the instantaneous vel-
ocity magnitudes can be calculated according to equa-
tion (3). Figure 3 is the instantaneous velocity
magnitude measured in case 5, including the velocity
in the area beside right arm (anemometer 1–6), in
front of right arm (anemometer 7–12), in front of left
leg (anemometer 13–18) and beside left leg (anemom-
eter 19–24). The data in Figure 3 lasted 10 s and began
at 30 s after the experiment started. In Figure 3, appar-
ent periodic motion of the airflow can be seen in the
area close to the wrist and ankle of the manikin, espe-
cially for the velocity measured by anemometer 6 and
23 located beside the arm and leg, respectively. From
Figure 3, the periodicity of the airflow motion was not
stable and was lagging behind the pendulum of the
limbs. When the limbs were swinging, the limbs may
have pushed the air in the pendulum direction away
and may have also drawn the air around the limbs to
follow the movement of the limbs. So, the induced air-
flow motion would follow and lag behind the pendulum
of the limbs. Because of the turbulence effects, the peri-
odicity of the airflow motion was unstable and not so
Figure 3. Instantaneous velocity magnitude of case 5: (a) beside right arm, (b) in front of right arm, (c) in front of left leg and
(d) beside left leg.
Han et al. 7
obvious in the further area from the limbs. Figure 3
also shows that the area close to the wrist (anemometer
6 and 12) and ankle (anemometer 18 and 23) has sig-
nificantly larger velocity than other areas. This larger
velocity may be because of larger swing amplitude of
the wrist and ankle. For the area close to the upper
arms (anemometer 1, 2, 7 and 8) and thigh (anemom-
eter 13, 14, 19 and 20), the velocity was almost zero,
due to the small swing amplitude of upper arms and
thigh.
Similar characteristics can also be found in cases
1–4. To demonstrate the effects of step frequency, com-
parison of the averaged velocity magnitudes around the
wrist and ankle of cases 1–5 is shown in Figure 4. Since
the motion of limbs pendulum was cyclical, the arith-
metic mean of the instantaneous velocity magnitude
measured in 60 s was used for this comparison. From
Figure 4, a steady rising of the average velocity magni-
tude was maintained with the increases of step fre-
quency. With the higher step frequency, a larger
velocity both for the area around the arm and leg had
resulted; the exceptions were measured by anemometer
6 in case 5. In case 5, the velocity of the airflow beside
the wrist was slightly smaller than that of case 4. That
was probably because the reciprocating motion of the
wrist may also have reduced the average velocity by
offsetting the original airborne in the opposite direc-
tion. Figure 4 also gives a small average velocity mag-
nitude (<0.1 m/s) in cases 1 and 2, indicating that the
limbs pendulum during strolling would not cause
apparent changes of the flow field in the nearby area.
In addition, as shown in Figure 4, the pendulum of
wrist and ankle may have induced a larger velocity on
the right side of wrist (anemometer 6) and in front of
the ankle (anemometer 18) than others, respectively.
Figure 5 gives the arithmetic mean of the instantan-
eous velocity magnitudes around the limbs with
different step frequencies. In the vertical direction,
larger velocity can be found in the area close to the
free end of the limbs (anemometer 5, 6, 11, 12, 17, 18,
23 and 24) than others due to larger swing amplitude of
the wrist and ankle. In the horizontal direction, the area
close to the limbs (anemometer 4, 6, 10, 12, 16, 18, 21
and 23) may also have larger velocity than others. For
the area close to the upper arms (anemometer 1, 2, 7
and 8) and thigh (anemometer 13, 14, 19 and 20), the
velocity was almost zero due to the small swing ampli-
tude of upper arms and thigh. By comparing the vel-
ocity measured around right arm and left leg, the
velocity beside the left leg was smaller than others.
These analysis results indicate that the flow field and
velocity distribution induced by limb pendulum around
the limbs are complicated. Significant airflow can be
seen around the wrist and ankle, and these would also
depend on the pendulum frequency and spatial
location.
Body motion
In this experiment, the temperature and relative humid-
ity distribution in the cabin did not change with the
movement of the manikin. During the experiment, the
temperature was 21.5–21.7C, and the relative humidity
was 23.7–25.2% with good stability.
By observing the velocity distribution measured in
repeated measurements, similarity and repeatability
were found. The results of instantaneous velocity mag-
nitudes are shown in Figure 6, corresponding to the hip
height of the manikin (anemometer 3, 9, 15, 21 and 27).
In Figure 6, apparent time relationship can be seen
between the instantaneous velocity magnitudes mea-
sured by different groups of anemometers. The velocity
measured by each group of anemometers remained zero
until the manikin passing by. So, the flow field in the
Figure 4. Comparison of the average velocity magnitudes of cases 1–5: (a) around right arm and (b) around left leg.
8Indoor and Built Environment 0(0)
environment could not significantly be affected by the
movement of the manikin unless the manikin arrived at
and passed that area. When the manikin went through
the location of each group of anemometers, the velocity
first increased slowly for a period of 0.3–0.5 s and then
increased rapidly to a peak velocity in less than 0.6 s.
After the manikin passed through that area, the vel-
ocity slowly decreased to around zero but remains
unstable. For the anemometers in group 1, it took
approximately 5–6 s for the velocity to decrease to
around zero. For group 2, it was nearly 8–10 s, and
for the other groups, it was even longer. The further
the manikin moved, the longer time it took for the flow
field to become static again. After each measurement, it
usually needed more than 2 min for the flow field in the
cabin to become static state again, which was much
longer than the time given by CFD simulations due
to the turbulence effects.
18
Higher moving speed of
the manikin had also lead to longer time for the flow
field to become stable again. So, the duration of the
influence of movement on flow field would depend on
moving distance and moving speed. From Figure 6, the
maximum velocity magnitudes measured at different
location along the moving path were different.
Especially, as shown in Figure 6(d) to (e), the maximum
velocity measured by anemometer 3 was significantly
smaller than the others. Figure 6 also shows that the
maximum velocity measured under different moving
speeds was different. So, the result indicates that the
aerodynamic effects of human movement would
depend on moving speed and moving distance.
Comparison of the results measured in cases 6–10 is
shown in Figure 7 to demonstrate the effects of moving
speed on the flow field. In Figure 7, the results are the
arithmetic mean of the maximum velocity magnitudes
measured by different groups of anemometer, for chest
(anemometer 7, 13, 19 and 25), hip (anemometer 9, 15,
21 and 27) and knee (anemometer 11, 17, 23 and 29),
respectively. By observing the relationship between the
moving speeds and the maximum velocity magnitude
shown in Figure 7, a monotonically increasing trend
can be found. Higher moving speed may have caused
airflow with a larger velocity. Figure 7 also shows that
the maximum velocities measured around chest and hip
are substantially equal and significantly larger than that
of knee. So, the torso may have generated a greater
effect on the flow field than the legs. This result indi-
cates that different parts of the human may also have
different aerodynamic effects on the flow field during
human movement due to the complexity of the shape of
human body.
To demonstrate the characteristics of velocity mag-
nitude distribution of the airflow induced by human
movement, Figure 8 gives the maximum velocity mag-
nitudes of all the anemometers. In the vertical direction,
larger velocity can be found around the chest and hip of
the human (anemometer 1, 3, 7, 9, 13, 15 19, 21, 25 and
27) than others. So, the profile of the manikin used in
this experiment also has a significant impact on the
airflow motion and velocity distribution of human
movement. When a human moves, more detailed pro-
file of the human body would lead to a more
Figure 5. Average velocity magnitude around arm and leg of cases 1–5 with different step frequencies.
Han et al. 9
complicated flow field along the moving path. In the
horizontal direction, the area close to the moving path
(anemometer with odd number) may also have larger
velocity than others. Farther area from the moving
path has smaller velocity. From Figure 8, the maximum
velocities measured by the anemometers of group 1
(anemometer 1–6) were smaller than those measured
by the anemometers of other groups (anemometer 7–
30), and the maximum velocities measured by the
anemometers at the same location of groups 2–5 were
similar to the results mentioned above. That is prob-
ably because larger moving speed may need longer
acceleration time, and longer moving distance would
be needed for the manikin to accelerate to the moving
speed. Especially for cases 9 and 10, the velocity of the
induced airflow following the movement behind the
human body was still increasing as the manikin passes
the location of anemometer 1–6. So, the manikin can
Figure 6. Instantaneous velocity magnitude of cases 6–10 at the height corresponding to the hip of the manikin: (a) case 6,
(b) case 7, (c) case 8, (d) case 9 and (e) case 10.
10 Indoor and Built Environment 0(0)
effectively induce the airflow along the motion path
after it has moved more than 4 m.
These analysis results indicate that the flow field and
velocity distribution induced by body motion around
the moving human are complicated. Significant airflow
can be seen around the hip and chest, which also
depends on the moving speed. So, the characteristics
of the velocity distribution induced by human move-
ment would depend on moving speed, moving distance
and spatial location.
Discussion
In this work, the influences of different human move-
ment behaviours on the flow field were compared and
analysed. The velocity measured around the same area
during different movement behaviours was compared,
as shown in Figure 9. In Figure 9(a), the maximum
velocity magnitude measured around the wrist of the
manikin (anemometer 6) of cases 1–5 was compared
with the average value of the maximum velocity mag-
nitude measured around the hip of the manikin along
the moving path (anemometer 9, 15, 21 and 27) of cases
6–10. In the vertical direction, these areas correspond
to the same height of the human body during human
walking. In Figure 9(b), the maximum velocity magni-
tude measured around the calf of the manikin
(anemometer 23) of cases 1–5 was compared with the
average value of the maximum velocity magnitude mea-
sured around the knee of the manikin along the moving
path (anemometer 11, 17, 23 and 29) of cases 6–10. In
the horizontal direction, the distances between the
anemometers and the human body were almost the
same in cases 1–10. From Figure 9(a), the velocity of
the airflow around the hip induced by body motion was
significantly larger than that around the wrist induced
by limbs pendulum. As shown in Figure 9(b), the vel-
ocity of the airflow around the knee induced by body
motion was also significantly larger than that around
the calf induced by limbs pendulum. In addition,
according to Figures 7 and 8, the velocity of the airflow
around the chest (measured by anemometer 7, 13, 19
and 25 in cases 6–10) was also larger than that around
the upper arm, which was almost zero (measured by
anemometer 1 and 2 in cases 1–5). So, the airflow
induced by body motion has much larger velocity
than that induced by limbs pendulum. This analysis
results indicate that body motion could impose a
strong influence on the flow field and is much greater
than that generated by the limbs pendulum movement.
The walking behaviours may significantly affect the
airflow motion in the nearby area around the human
body when a human walks. For limbs pendulum, the
swings of the arms could cause leakage airflow between
the torso and the arms, and the swings of the legs could
also cause leakage airflow among the legs. These leak-
age airflows would move backward and forward
through the gap between the torso, arms and legs and
could result in more sufficient mixing of the air around
the human body. For the body motion, the moving
human can push the air in front of the human
upward and sideward bypassing the human body. The
bypassing airflow would move backward and forward
to the two ends of the cabin. Vortex flow would occur
Figure 7. Comparison of the maximum velocity magnitude of cases 6–10.
Han et al. 11
around the bordering of the human body, similar with
the flow around the free end of finite cylinder in free
stream.
29
After the human passes through, the bypass-
ing airflow would be drawn and move towards the
moving path and would gradually combine with the air-
flow following the human movement behind the human
body. So, in the horizontal direction, two symmetrical
recirculation regions would be formed around the
human body, which may have promoted the airflow
motion in the moving direction and enhanced the air
mixing in the cabin.
During human walking, the wake behind the human
body also has a strong influence on the flow field in the
cabin. Asymmetric vortex shedding could exist in an
unsteady wake behind the human body, and the aero-
dynamic effects could also depend on the moving speed
and spatial location.
16
In this work, the complicated
profile of the manikin may have also resulted in a com-
plicated unsteady wake, and the wake behind different
parts of the human body may also have different char-
acteristics. According to the previous literatures on the
two-dimensional flow field that two circular cylinders
Figure 8. Maximum velocity magnitude of the anemometers of cases 1–5 with different moving speeds.
Figure 9. Comparison of the influence of body motion and limbs pendulum on the flow field for cases 1–10: (a) around wrist
and hip and (b) around calf and knee.
12 Indoor and Built Environment 0(0)
could be configured side-by-side, three different regimes
of the wake structure would be found, which rely on the
gap width between the two cylinders.
30,31
The first
regime of wake would be a combined wake behind
the circular cylinders combined together and no sepa-
rated wake exists. This kind of wake occurs when the
cylinders are placed closely together, and the gap width
between them is smaller than 20% of the diameter of
the cylinders. So, the wake behind the upper arms and
the chest of the manikin is corresponding to this
regime, and a large combined wake would be formed
behind the back of the manikin. This wake is also much
larger than the wake behind other parts of the manikin
because of the large width of the torso. For the second
regime of wake, individual wake can be found behind
each circular cylinder, which is also heavily influenced
by the neighbouring wake. This kind of wake occurs
when the gap width between the cylinders ranges from
20% to 120% of the cylinder diameter. So, this kind of
wake may exist behind the chest, waist, hip and arms. If
the gap width between the cylinders is larger than 120%
of the cylinder diameter, independent wakes are
formed, which correspond to the third regime of
wake. This kind of wake can be found among the
legs, knees and ankles. So, by comparing with the pre-
vious literatures, these results suggest that the profile of
the human body has a strong influence on the aero-
dynamic effects of human movement, especially in the
wake behind the human body.
In indoor environment, these aerodynamic effects of
human walking can also enhance the pollutant disper-
sion and transportation. Different human movement
behaviours may have different effects on the generation
and dispersion of the contaminants. During human
activities, the human body with the clothes that cover
it may be an important source of indoor contamin-
ants.
32
A large number of particles can be emitted
from the human body during limbs pendulum because
of the friction of clothes and the induced airflow
around the human.
33
As shown by Edge et al.,
16
the
pollution source was defined on the surface of the
human body, and the effects of vortex shedding in the
wake recirculation region would result in much higher
level of contaminants in the wake behind the human
body. The airflow between the arms and the torso as
well as the legs could affect the turbulence in the wake
and result in the diversification of contaminants con-
centration distribution.
16
So, when a human walks, the
human body may become a moveable source of pollu-
tant because of the friction of clothes during limbs pen-
dulum. The particle concentrations along the moving
path may also increase due to the aerodynamic effects
of limbs pendulum. Different from limbs pendulum, the
body motion during human walking may result in the
re-suspension of the particles from the flooring, which
is also an important source of particulate matters in
indoor environment and could lead to a significant
increase of particle concentration around the human
body.
34
As demonstrated by numerical simulations in
room–room, room–hall and compartments–vestibule–
compartments configuration, the human-induced
wake motion can also enhance the contaminant trans-
port when a person walks from a contaminated region
to a clean region.
20,35
The wake-induced contaminants
transport in the walking direction may continue due to
inertial effects even after the human stops. Similar pro-
motion effects can also be found in isolate room and
airplane cabin.
19,22
So, the body motion can increase
the particle concentration in the nearby area, and its
aerodynamic effects may also significantly enhance the
transport of pollutants and increase the personal expos-
ure to indoor contaminants of the occupants.
According to the studies on the wake behind finite
cylinder in free stream, the distance needed for the
wake and turbulence to sufficiently develop in three-
dimension is about five times of the diameter of the
cylinder.
29,36
For the manikin used in this experiment,
the distance was at least 1.45 m away behind the human
body. For the experimental setup of body motion in
this work, it was difficult to install the velocity probes
on the trolley 1.45 m away behind the human body and
let them move with the manikin due to the limitation of
the hot-wire anemometers. So, the velocity distribution
in the wake behind the human body was not measured
and compared in this experiment. For future studies on
the velocity distribution and turbulence characteristics
of the airflow induced by human movement, large-scale
and non-contact measurement technology or moveable
measurement instrument would be strongly needed.
CFD simulation would also be an important and neces-
sary method for the verification of the measured results
and for the investigation on the turbulence characteris-
tics of the induced airflow.
Conclusions
In this work, the aerodynamic characteristics of human
movement behaviours were investigated. The velocity
of the airflow induced by limbs pendulum and body
motion was measured, and the characteristics of the
induced flow field were analysed. Similarity and repeat-
ability were found in the results of the repeated meas-
urements. The experimental results of limbs pendulum
indicate that the characteristics of the velocity distribu-
tion induced by limbs pendulum would depend on pen-
dulum frequency and spatial location. The airflow
caused by wrist and ankle pendulum has larger velocity
than upper arm, forearm, calf and thigh. Higher step
frequency could also induce airflow with larger velocity,
both for areas around arm and leg. For body motion,
Han et al. 13
the experimental results indicate that the characteristics
of the velocity distribution induced by human move-
ment would depend on moving speed, moving distance
and spatial location. The manikin can effectively induce
the airflow along the motion path after it has moved
more than 4 m. Significant airflow can be seen around
the hip and chest. Higher moving speed of the manikin
could lead to longer period of time for the flow field to
become stable again. Farther areas away from the
manikin motion path have smaller velocity. The move-
ment of the manikin and the induced airflow would not
affect the temperature and relative humidity distribu-
tion in the experimental area.
By comparing the velocities of the induced airflow of
different human movement behaviours, the body
motion has a significant influence on the flow field,
and its aerodynamic effects would be much greater
than that of limbs pendulum. When a human walks,
the wake behind the human body would pose a
strong influence on the flow field in the cabin. The aero-
dynamic effects of the complicated wake may signifi-
cantly enhance the pollutant dispersion and
transportation and should be investigated in the
future work.
Acknowledgements
This paper was supported by China National Key Basic
Research Special Funds Project (Grant No. 2012CB719705),
National Natural Science Foundation of China (Grant No.
51076073 and 91024018), Tsinghua University Initiative
Scientific Research Program (Grant No. 2012THZ02160)
and National Twelve Five-Year Scientific and Technical
Support Plans (2011BAK07B03).
References
1. WHO. International travel and health. Geneva: World Health
Organization, 2007, p.74.
2. WHO. The World health report. Geneva: World Health
Organization, 2004, pp.121–123.
3. Gupta JK, Lin CH and Chen Q. Flow dynamics and character-
ization of a cough. Indoor Air 2009; 19: 517–525.
4. Moser MR, Bender TR, Margolis HS, Noble GR, Kendal AP and
Ritter DG. An outbreak of influenza aboard a commercial air-
liner. Am J Hyg 1979; 110: 1–6.
5. Klontz KC, Hynes NA, Gunn RA, Wilder MH, Harmon MW and
Kendal AP. An outbreak of influenza A/Taiwan/1/86 (H1N1)
infections at a naval base and its association with airplane
travel. Am J Epidemiol 1989; 129: 341–348.
6. Kenyon TA, Valway SE, Ihle WW, Onorato IM and Castro KG.
Transmission of multidrug-resistant mycobacterium tuberculosis
during a long airplane flight. N Engl J Med 1996; 334: 933–938.
7. Miller MA, Valway S and Onorato IM. Tuberculosis risk after
exposure on airplanes. Tuber Lung Dis 1996; 77: 414–419.
8. Olsen SJ, Chang H-L, Cheung TY-Y, Tang AF-Y, Fisk TL, Ooi
SP-L, Kuo H-W, Jiang DD-S, Chen K-T, Lando J, Hsu K-H,
Chen T-J and Dowell SF. Transmission of the severe acute respira-
tory syndrome on aircraft. N Engl J Med 2003; 349: 2416–2422.
9. Wilder-Smith A, Leong HN and Villacian JS. In-flight transmis-
sion of Severe Acute Respiratory Syndrome (SARS): a case
report. J Travel Med 2003; 10: 299–300.
10. Leder K and Newman D. Respiratory infections during air
travel. Intern Med J 2005; 35: 50–55.
11. Mangili A and Gendreau MA. Transmission of infectious dis-
eases during commercial air travel. Lancet 2005; 365: 989–996.
12. DeHart RL. Health issues of air travel. Annu Rev Publ Health
2003; 24: 133–151.
13. Gendreau M. Tuberculosis and air travel: guidelines for preven-
tion and control, 3rd edition. Perspect Public Health 2010; 130:
191–191.
14. McFarland JW, Hickman C, Osterholm MT and Macdonald
KL. Exposure to mycobacterium-tuberculosis during air-travel.
Lancet 1993; 342: 112–113.
15. Mazumdar S, Poussou SB, Lin CH, Isukapalli SS, Plesniak MW
and Chen QY. Impact of scaling and body movement on con-
taminant transport in airliner cabins. Atmos Environ 2011; 45:
6019–6028.
16. Edge BA, Paterson EG and Settles GS. Computational study of
the wake and contaminant transport of a walking human.
J Fluids Eng 2005; 127: 11.
17. Shih YC, Chiu CC and Wang O. Dynamic airflow simulation
within an isolation room. Build Environ 2007; 42: 3194–3209.
18. Mazumdar S, Yin YG, Guity A, Marmion P, Gulick B and Chen
QY. Impact of moving objects on contaminant concentration
distributions in an inpatient ward with displacement ventilation.
HVAC&R Res 2010; 16: 545–563.
19. Wang J and Chow T-T. Numerical investigation of influence of
human walking on dispersion and deposition of expiratory drop-
lets in airborne infection isolation room. Build Environ 2011; 46:
1993–2002.
20. Choi JI and Edwards JR. Large-eddy simulation of human-
induced contaminant transport in room compartments. Indoor
Air 2012; 22: 77–87.
21. Matsumoto H and Ohba Y. The influence of a moving object on
air distribution in displacement ventilated rooms. J Asian Archit
Build Eng 2004; 3: 71–75.
22. Poussou SB, Mazumdar S, Plesniak MW, Sojka PE and Chen
QY. Flow and contaminant transport in an airliner cabin induced
by a moving body: model experiments and CFD predictions.
Atmos Environ 2010; 44: 2830–2839.
23. Bjorn E and Nielsen PV. Dispersal of exhaled air and personal
exposure in displacement ventilated rooms. Indoor Air 2002; 12:
147–164.
24. Thatcher TL, Wilson DJ, Wood EE, Craig MJ and Sextro RG.
Pollutant dispersion in a large indoor space, part 1: scaled experi-
ments using a water-filled model with occupants and furniture.
Indoor Air 2004; 14: 258–271.
25. Gao NP and Niu JL. CFD study of the thermal environment
around a human body: a review. Indoor Built Environ 2005; 14:
5–16.
26. Bruun HH. Interpretation of hot-wire probe signals in subsonic
airflows. J Phys E: Sci Instrum 1979; 12: 1116–1128.
27. Jørgensen FE. Directional sensitivity of wire and fibre-film
probes. DISA Inf 1971; 11: 31–37.
28. Fitouri A, Khan MK and Bruun HH. A multiposition hot-wire
technique for the study of swirling flows in vortex chambers. Exp
Therm Fluid Sci 1995; 10: 142–151.
29. Krajnovic S. Flow around a tall finite cylinder explored by large
Eddy simulation. J Fluid Mech 2011; 676: 294–317.
30. Meneghini JR, Saltara F, Siqueira CLR and Ferrari JA.
Numerical simulation of flow interference between two circular
cylinders in tandem and side-by-side arrangements. J Fluids
Struct 2001; 15: 327–350.
14 Indoor and Built Environment 0(0)
31. Sumner D, Wong SST, Price SJ and Paidoussis MP. Fluid behav-
iour of side-by-side circular cylinders in steady cross-flow.
J Fluids Struct 1999; 13: 309–338.
32. Wallace LA, Emmerich SJ and Howard-Reed C. Effect of central
fans and in-duct filters on deposition rates of ultrafine and fine
particles in an occupied townhouse. Atmos Environ 2004; 38:
405–413.
33. You R, Cui W, Chen C and Zhao B. Measuring the short-term
emission rates of particles in the ‘‘personal cloud’’ with different
clothes and activity intensities in a sealed chamber. Aerosol Air
Qual Res 2013; 13: 911–921.
34. Xinyu Z, Ahmadi G, Jing Q and Ferro A. Particle detachment,
resuspension and transport due to human walking in indoor
environments. J Adhes Sci Technol 2008; 22: 591–621.
35. Choi JI and Edwards JR. Large Eddy simulation and zonal mod-
eling of human-induced contaminant transport. Indoor Air 2008;
18: 233–249.
36. Park CW and Lee SJ. Flow structure around a finite circular
cylinder embedded in various atmospheric boundary layers.
Fluid Dyn Res 2002; 30: 197–215.
Han et al. 15