Building and Environment 77 (2014) 101e109

Contents lists available at ScienceDirect

Building and Environment

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

Human induced flow field and resultant particle resuspension and

transport during gait cycle
Iman Goldasteh a, Yilin Tian b, Goodarz Ahmadi a, *, Andrea R. Ferro b
a
Department of Mechanical and Aeronautical Engineering, Clarkson University, NY, USA
b
Department of Civil and Environmental Engineering, Clarkson University, NY, USA

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

Article history: Particle resuspension from flooring is believed to be an important source of particulate matter (PM) in
Received 2 December 2013 the indoor environment. It is hypothesized that the high speed airflow generated between the flooring
Received in revised form and the foot during the gait cycle is the main cause of particle resuspension. The simulation results show
15 March 2014
that particles are detached from the substrate during the downward motion of the foot. Furthermore,
Accepted 18 March 2014
Available online 26 March 2014
during the upward motion of the foot, additional particles may also be resuspended because of the
suction flow generated by lifting the foot. These resuspended particles are then dispersed in the room by
the airflow turbulence.
Keywords:
Particle resuspension
In this study a combined experimental and computational studies were performed to shed light on the
Indoor air quality mechanisms of particle resuspension from flooring during the gait cycle. A mechanical foot experimental
Turbulent flow setup which mimics the human walking was developed and used to measure the rate of resuspension. In
Particle dispersion addition, to investigate the airflow field during walking, a three dimensional numerical model of a
CFD moving shoe during the gait cycle was generated using the ANSYS-FLUENTÔ CFD package. A RANS
Gait cycle approach with the RNG k-epsilon turbulence model was used for simulating the unsteady airflow field
around and under the shoe. To include the shoe motion in the analysis, a user defined function (UDF) was
developed and the dynamic mesh technique was used. A resuspension model was applied for the
resuspension of pre-deposited particles from the flooring. The predicted particle resuspension were
compared with the corresponding experimental results and good agreement was found.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction indoor activities, walking is recognized as a major contributor to

Particle resuspension is known to be an important source of
particulate matter in the indoor environments [1]. Resuspension is
defined as a process in which the pre-deposited particles on the
floor are detached under the effects of fluid flow, surface acceler-
ation, and/or electrostatic force and the particles enter the flow
stream over the surface.
People spend approximately 90% of their time indoors [2]. The
rise in occurrence of allergic diseases has been associated with the
increased exposure to the indoor allergens including dust, pollen,
viruses, bacteria, fungi and various chemical agents [3]. Several
studied showed that indoor activities like vacuuming, walking, and
using upholstered furniture can increase the PM concentration to
orders of magnitude higher than the background level [4]. Among
* Corresponding author.
E-mail addresses: goldasi@clarkson.ed (I. Goldasteh), ahmadi@clarkson.edu
(G. Ahmadi).
the high PM concentration in indoor environments. Studies have
shown that walking in a room can increase the concentration of
particulate matter as much as 100% for submicron particles [1,5].
Walking also leads to even larger increase in indoor PM concen-
tration for micron range particles, especially 5e10 mm particles
[6,7].
Particle resuspension can be studied through both large-scale
bulk resuspension and small-scale detailed mechanism [8]. Large-
scale, or bulk, resuspension studies have expressed resuspension
in terms of the resuspension factor (particle concentration in air
divided by particle concentration on the surface), resuspension rate
coefficient (flux of particles joining the flow field over the surface
divided by particle concentration on the surface), or resuspension
fraction (fraction of particles on the surface resuspended into the
air per foot step). Large-scale resuspension studies are usually
motivated for predicting the resulting bulk airborne PM concen-
trations and related human exposures [9]. On the other hand,
small-scale resuspension studies include theoretical and experi-
mental studies on the details of resuspension process and

http://dx.doi.org/10.1016/j.buildenv.2014.03.016
0360-1323/Ó 2014 Elsevier Ltd. All rights reserved.
102 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109
determining the removal mechanisms of particles under the effects
of removal forces overcoming the adhesion forces. As discussed
above, a number of numerical and experimental works were con-
ducted during recent years to study the small-scale resuspension
phenomenon [10e19]. Generally, micron and submicron sized
particles in contact with a surface experience a strong van der
Waals adhesion force. Several experimental studies [12,13,20e22]
on removal and resuspension of particles from substrates showed
that a shear velocity, on the order of 1 m/s or larger, is required to
detach and resuspend the particles into the airflow field. This range
of shear velocity is larger than that associated with the movement
of the human body and shoe motion during the gait cycle far from
the flooring. However, experimental studies [1,5,23] showed that
human walking results in dust resuspension and a large increase in
the indoor PM concentrations. Earlier efforts hypothesized that the
high velocity jet ejected from the perimeter of an impacting object
during descending movement causes the particle detachment and
resuspension [24,25]. Madler and Koch [24,25] used a simplified
model to describe the particle resuspension caused by the radial
wall jet generated from an impacting disk to a dusty surface. Based
on their work, several authors developed models for estimating the
particle resuspension during gait cycle. Khalifa and Elhadidi [26]
used an analytical solution for modeling unsteady flow resulting
from the downward motion of a circular disk toward a seeded
substrate. They also performed a computational fluid dynamics
(CFD) analysis for the wall jet spreading radially outward of the disk
perimeter and investigated the effects of airflow on detachment of
spherical micron-sized particles.
Zhang et al. [27] studied the particle detachment, resuspension
and transport due to human walking using analytical and computer
simulation approaches. They modeled the stepping motion of the
foot as a downward and upward motion. They simplified the shoe
geometry with two circular disks for the toe and heel parts. They
modeled the squeeze film flow between circular disks and the floor
and evaluated the corresponding velocity field. They modeled the
airflow outside the foot perimeter as a wall jet. For simulating the
resuspended particle trajectories they used the Lagrangian
approach. The dispersion of resuspended particle clouds were also
included in their model. Kubota et al. [28] studied the human foot
movement visually and through particle image velocimetry (PIV)
measurements. Following Khalifa and Elhadidi [26] and Zhang et al.
[27], they simplified the human walking motion with a disk moving
normal to the floor downward and upward. Flow visualization of
the resuspension of pre-deposited particles on the flooring showed
particles are resuspended by both downward and upward motion
of the disk. For both the downward and upward motions, the high
speed wall jet emanating from the perimeter of the disk and
flooring was determined as the cause for resuspension. They also
determined that large scale ring vortex structures formed during
both downward and upward motions. While these vortices may not
contribute substantially to particle detachment process, they are
certainly highly effective at dispersing and the resuspended parti-
cles away from the floor.
Using particle image velocimetry (PIV), Eisner et al. [29] studied
the transport and dispersion of dust resuspended by a mechanical
foot apparatus. They showed that during the motion of the shoe
when the heel detaches and shoe rotates around a point in the ball
of the foot, a draft corner flow develops, which carries the particles
from the region heel to toe area. Oberoi et al. [30] modeled particle
resuspension from the up and down motion of the foot using an
Eulerian approach. Choi and Edwards [31,32] and Choi et al. [31,32]
used the same procedure and modeled the removal of particles
during gait cycle. Recently Kubota and Higuchi [33] experimentally
investigated particle resuspension due to human foot motion
focusing on aerodynamic effects using PIV.
Previous studies have reported important findings on resus-
pension during the walking process. Although a number of large-
scale and small-scale studies have been reported, there is a gap
between these two fields where current theoretical resuspension
models do not directly predict the particle number concentration
for an indoor activity, such as walking considering detail interaction
of particle in contact with the substrate. In this study, a reliable
model of particle resuspension during walking activity which is not
available with details as yet was introduced. In this work, a com-
bined experimental and computational study was performed to
shed light on the mechanisms of particle detachment from flooring
during the gait cycle and the corresponding resuspension and
dispersion in indoor environment. A mechanical foot experimental
setup was developed and used to measure the rate of resuspension
for a range of conditions. The mechanical foot device mimics hu-
man walking and may be used to generate reproducible data for
particle resuspension. In addition, a three dimensional numerical
model of a shoe during heel-to-toe walking activity was generated
using the ANSYS-Fluent computational fluid dynamic (CFD) pack-
age. The motion of the shoe was modeled using UDFs in the CFD
software. The particle trajectories were analyzed using the
Lagrangian approach including drag, lift and gravity forces. The
Monte Carlo simulation model developed by Goldasteh et al. [10]
was used for analyzing the detachment of particles from the floor.
This new model takes into account the effects of dust particle ir-
regularity, dust particle surface roughness, size distribution of
particles and the spatial orientation of each dust particle on the
flooring. With these modifications, the general model predicted the
resuspension process during gait cycle more realistic than previous
models which agreed well in compare to the experiments.
2. Experimental study
Experimental study of particle resuspension due to the gait cycle
was conducted using a mechanical shoe resuspension apparatus, as
shown in Fig. 1. This setup was designed to mimic the human foot
step during walking. The resuspension apparatus consists of toe
and heel plates controlled by two electric actuators. A men U.S. size
ten tennis shoe was mounted on the resuspension apparatus and
used in these experiments. Tian et al. [34] have reported that the
resuspension apparatus provided consistent rotation motion,
stepping frequency and pressure loading, which are comparable to
human foot step. The mechanical show apparatus was housed in a
Fig. 1. Mechanical foot setup in the resuspension chamber.
 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109
61
38
58 cm Plexiglas resuspension chamber with a volume of
0.12 m3. The chamber was sealed during each experiment, with
temperature and relative humidity controlled at 23  0.3  C and
24  0.02%, respectively.
Vinyl sheet flooring with a ToughGuardÒ wear layers composi-
tion and a CleanSweepÒ no-wax wear surface was used in the
experimental study. Flooring samples were cut into 0.15
0.3 m
piece and then soiled with 6 g of ATD using the reproducible
seeding system described by Tian et al. [34,35], resulting in a sur-
face dust loading of 2.16 g/m2. The post-seeding particle volume
size distribution of ATD was estimated by the Malvern Mastersizer
2000 laser diffraction particle size analyzer (Worcestershire, UK).
The volume fractions of 5e7.5 mm and 7.5e10 mm particles were,
respectively, 22% and 7% of the total dust.
To measure the real-time size-resolved airborne concentration
at the toe and side of the shoe, two Grimm Technologies (Dou-
glasville, GA) model 1.108 portable laser aerosol spectrometers
(Grimm) were used. For 5e7.5 mm and 7.5e10 mm size bins, relative
precision of the two Grimm monitors after adjustment based on
collocation results was 14% and 25%, respectively. A HOBO H08 data
logger (Onset Computer Corporation, Waltham, MA) was installed
in the resuspension apparatus to monitor the temperature and
relative humidity. The schematic diagram of the experimental
setup is shown in Fig. 2. Filtered air was supplied from the top of the
resuspension chamber at 2.8 L/min to make up for the air loss due
to the sampling instruments. Walking speed was adjusted to 0.5
step/s and shoe started the gait cycle from the horizontal position
at the elevation of 0.086 m and returned to the initial position after
2 s. The gait cycle consists of the following steps: A downward
motion of the heel when the heel section arrives to the substrate
level, following by the same motion for the toe. At the end of this
step shoe is placed on the surface. Similarly, for the upward motion,
there is an upward motion of the heel section following by the same
motion for the toe. The details of the resuspension apparatus and
Fig. 2. Schematic diagram of the resuspension chamber setup. Optical particle coun-
ters Grimm 1 and Grimm 2 sampling locations: 5 cm above the floor (both monitors)
and 5 cm away from the center point of the long (side part) and short edge (toe part) of
the flooring, respectively.
103
Table 1
Procedure which is used in experimental study.
Time (min) Activity
0e5 Clean up the shoe bottom and resuspension chamber
5e10 Place the flooring sample; seal the resuspension chamber;
turn on the two Grimms
10e100 Background period (resuspension chamber undisturbed)
100e102 Operate the resuspension apparatus for 30 continuous steps
102e140 Decay period (resuspension chamber undisturbed)
specifications were included in the thesis by Tian [36]. Three
replicated experiments were conducted following the procedures
listed in Table 1.
3. Numerical simulation
3.1. Numerical scheme
To numerically investigate the effect of walking on particle
resuspension from flooring, a three dimensional model of a moving
shoe during the gait cycle was developed. The geometry of the shoe
is shown in Fig. 3. A number of meshes were tested and grid
independency was obtained with a non-uniform tetrahedral mesh
of 668,736 elements. The movement of the shoe and the corre-
sponding grid deformation were included using a User Defined
Function (UDF) in the ANSYS-Fluent code along with re-meshing
methods available in the CFD package [37]. Similar to earlier
studies [27], it is assumed that when the foot reaches the flat po-
sition, there is a 0.25 mm gap between the shoe and flooring due to
the grooves at the bottom of the shoe.
3.2. Governing equations
3.2.1. Flow field
For modeling the unsteady, incompressible flow field during the
gait cycle, the Reynolds averaged NaviereStokes equations (RANS)
was employed. For closing the system of governing equations, the
Reynolds stress tensor must to be evaluated the use of appropriate
turbulence model. Here, RNG k-epsilon turbulence model was used
because of the simplicity, robustness and lack of computational
efforts for simulating the airflow field around the shoe. As this
approach fails the near wall region, a standard wall function was
implemented to resolve the problem near the boundaries [38]. The
governing equation can be expressed as the general form as,
v
! 
ðr4Þ þ div r u 4  G4 grad4 ¼ S4 (1)
vt
Fig. 3. Geometry of the shoe model and the computational grid used.
104 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109

Fig. 4. Variations of velocity magnitude contours (m/s) during downward and upward motions of the shoe near a horizontal flooring at different times (s).

Here, 4 represents each of the three velocity components u, v, w

and turbulent kinetic energy, k, and dissipation rate, ε. G4 is the Cc ¼ 1 þ Kn½1:257 þ 0:4 expð1:1=KnÞ (4)
effective exchange coefficient for the variable 4 and S4 is the source
term of the equations [39]. Here Kn is the Knudsen number which is defined as,

2l
3.2.2. Discrete phase model Kn ¼ (5)
d
Tracking particles was performed with a Lagrangian approach
by solving the equation of particle motion. Considering the particle where l is the mean free path of the fluid.
size and density in indoor environment some of the forces are small
enough and could be neglected. The detailed analysis on choosing
3.2.3. Boundary conditions
the proper external forces on particles was given by Zhao et al. [40].
Numerical simulation was performed for the same geometry
Here the discrete random walk (DRW) model is used in the analysis
which was explained in detail in Section 2. The inlet filtered air
to account for the effect of instantaneous turbulent velocity fluc-
which was supplied from the top of the resuspension chamber was
tuations on particle dispersion. The effects of inertia, viscous drag,
defined as a section with uniform velocity inlet and the measuring
Saffman lift and gravity were included and other forces such as
tubes was considered as the pressure outlet boundary conditions.
pressure gradient force, basset force and virtual mass force were
For particles also the trap type boundary condition was applied
neglected [40]. The corresponding particle equation of motion can
for the Plexiglas chamber. In the case of outlet flow, the escape type
be express as,
boundary condition was applied to mimic the same situation as
   experiments.
CD Re  
p
dui 18m
Ui þ u0i  ui þ gi þ Fl
p
¼ (2)
dt rp d2 Cc 24
3.3. Particle detachment and resuspension model
and
Indoor dust particles are typically compact non-spherical with a
dxpi p
wide size distribution and highly rough surfaces, which makes
¼ ui (3) modeling their adhesion to surfaces more challenging. In this work,
dt
a newly developed removal model for irregularly shaped particle
Here, upi is particle velocity, xpi is particle position, d is diameter, m is was employed. Earlier, Soltani and Ahmadi [44] developed the
dynamic viscosity, CD is drag coefficient, Re is relative Reynolds bumpy particle model for resuspension of compact irregularly
number, gi is acceleration of gravity, Fl is the Saffman lift force per shaped particles from substrates. In this approach, an irregularly
unit mass of particles [41], rp is particle density, and Cc is the shaped particle is modeled as a sphere with a number of bumps.
Cunningham factor which is given by Fuchs [42] and Friedlander Recently, Goldasteh et al. [10] included the effects of surface
[43] as, roughness and its variability, particle size distribution, and
 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109 105

Fig. 5. Variation of shear velocity (m/s) on the surface during gait cycle versus time (sec). (Gap size ¼ 0.25 mm, walking speed ¼ 0.5 step/s).

randomness of bump orientation with respect to the flow direction number of bumps, K represents the composite Young’s modulus, n
in the model. Here, the rough bumpy particle model with three represents the number of asperities per unit area, Wa is the ther-
contact bumps is used. Accordingly, the particles are detached from modynamic work of adhesion between particle and surface, br, is
the substrate when the contact of upstream rough bump breaks the mean curvature of roughness asperities, and Dc is a non-
from the substrate and particle rotates around the axis formed with dimensional roughness parameter that represents the ratio of
the two other bumps through the so-called rolling detachment elastic displacement of tip of each asperity dc at the moment of
mechanism. At the moment of removal the adhesion force of each detachment from substrate and the standard deviation of
rough bump to the surface can be expressed as [10,45], roughness heights s, which introduced by Fuller and Tabor [46]
which is
!2
d h i3
1:5p2 nWa br exp½  0:6=ðD2c Þ
Rough dc
Fpo ¼ pffiffiffiffi (6) Dc ¼
nu nb N K (7)
s
Here, d is the diameter of the particle, nb ¼ 1,2,. is the average where dc is given as,
spacing between bumps which in this work nb ¼ 2 is used that
!13
assumes there is one spacing between bumps, nu is a constant 2
fpo
greater than 1 that allows for non-uniform distribution of bumps dc ¼ (8)
3K 2 b
on the particle surface with nu ¼ 1 corresponding to the uniform
distribution of bump spacing and used in this work, N is the
Table 3
Table 2 Material Properties [16].
Chemical composition of ATD particles used in the experiments.
Particle-substrate E1  E2 Wa y1y2
Component SiO2 Al2O3 Fe2O3 Na2O3 CaO MgO TiO2 K2O (109 N/m2) (J/m2)

Weight (%) 68e75 10e15 2e5 2e4 2e5 1e2 0.5e1.0 2e5 ATD-Vinyl 70e2.8 0.0174 0.16e0.41
106 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109

Table 4 roughness significantly reduces the real contact area, the pull-off
Statistical properties of various parameters. force given by Equation (6) is typically much smaller than that
Parameters Distribution Mean value Standard given by the JKR model for a smooth contact. Here, it is assumed
deviation that if the non-dimensional roughness parameter Dc becomes very
Particle diameter (d) Log-normal 6.75 mm, 8.25 mm 1.5 mm large, which is the case for a smooth surface, the maximum value of
Non-dimensional Log-normal 1.21 0.1 pull-off force obtained by Equation (6) approaches the value pre-
roughness (1/DC) dicted by JKR adhesion theory. Based on this assumption, the
Number of asperities Poisson Equation (11) e
number of roughness asperity per unit area is given as,
per unit area n
pffiffiffiffi!1=3
0:00145 K 2 nu nb N
n ¼ 3
(11)
Here, composite Young’s modulus, K, depends on the particle dWa2 br
and surface properties [10] and fpo is the adhesion force for each
asperity based on the JKR adhesion model. That is, Using force balance, rolling detachment mechanism and burst/
inrush turbulence model, the shear velocity needed for resus-
3 pension particles as a function of particle, surface and flow char-
fpo ¼ pWa br (9)
2 acteristic parameters can be expressed as [49],
It should be pointed out that the root mean square of asperity pffiffiffi !1=2
heights s can be measured based on roughness profiles of particle 3Fpob b nb Cc
u*Burst=Inrush ¼

 
and substrate surfaces through atomic force microscopy (AFM). 1:72prd3 CosðaÞ 0:75f 1þ0:15Re0:678 þfm
Measuring the asperity radius, br however, is quite challenging.
(12)
Rabinovich et al. [47,48] extensively investigated the effect of
nanoscale roughness at the contact area using AFM and developed where r is the fluid density, f ¼ 1.7009 and fm ¼ 0.943993 are
a model which describes the asperity radius b based on the correction factors for the wall effect given by O’Neil [50].
measurable parameters of the root mean square of asperity height, This particle detachment model was employed in the CFD
s. The radius of the asperity from the proposed model could be simulation of the airflow and particle dispersion due to shoe
estimated for the case of nanoscale roughness as, movement through several UDFs in the ANSYS-Fluent CFD pack-
age [37]. In these simulations, particle diameter follows a
br ¼ 1:485s (10)
lognormal distribution described by the measured mean diameter
Also, the number asperity per unit area varies based on the and standard deviation of the samples from the experimental
surface material and its roughness profile. Since the surface study. The number of bumps and asperities per unit area follow

Fig. 6. Sample particle dispersion patterns at different times during one stepping cycle from upside view (shoe geometry in hidden for better view).
 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109
the Poisson distributions based on scanning electron microscopy
(SEM) photos of particles [49] and the mean values obtained from
Equation (11). The non-dimensional roughness parameter follows
a lognormal distribution as well [49]. The critical shear velocity
for resuspension of different particles from substrate are evalu-
ated and used for triggering the particle removal from the sub-
strate in the CFD simulation. That is, when the simulated shear
velocity on the surface reaches or exceeds the critical value at the
location of a particle, the particle is assumed to be detached from
the surface, and its trajectory is tacked using the particle equation
of motion.
3.4. Initial particle loading and particulate matter (PM)
concentration
In this work, a MATLAB program was developed for generating a
population of random sizes for particles based on their size distri-
bution reported in the experimental study. The particles then are
distributed randomly on the test section with the use of a UDF in
the ANSYS-Fluent package [37]. That is, the simulation conditions
are the same as the experimental conditions for particle size dis-
tribution and mechanical shoe motion. Due to the large number of
particles less than 10 mm in the experiment, simulating the same
particle surface loading is not possible. Therefore, each particle in
the simulation represents a collection of experimental particles
which follow a nearly identical path. That is, the representative
particles in these simulations show the mean behavior of the
dispersed phase in the generated flow field. In the experiments, A1
Ultrafine Test Dust (ATD, Powder Technology Inc., Burnsville, MN)
particles were used. In these simulations, particles in the size
ranges of 5e10 mm were picked from a population with the
measured size distribution of ATD. An experimental surface loading
of 2.16 g/m2 was selected for these simulations, leads to approxi-
mately 602,334,000 particles in the size range of 5e10 mm on the
surface. In the simulations, 602,334 particles were distributed on
the floor section. That is, each simulated particle is representative
Fig. 7. Sequence of particle detachment and dispersion during the gait cycle. Snap shots are shown after 4th stepping cycle.
107
of the 1000 particles in the experiment which follow roughly the
same trend and mean path. As noted before, the Lagrangian particle
tracking approach was used for evaluating the particle trajectories
after their detachment form the surface. The two particle monitor
inlet locations near the toe part and side part of the foot were
modeled in the CFD simulation. Particle number crossing the inlet
locations was obtained versus time.
4. Results and discussion
4.1. Induced flow field during gait cycle
Fig. 4 shows the sequence of walking induced airflow velocity
magnitude on a vertical plane during stepping motion versus time.
Here, walking speed is 0.5 step/s, which is identical to that in the
experiment where shoe starts from the horizontal position at t ¼ 0
and return to the initial position after 2 s. This figure shows the
formation of high velocity region under the shoe sole near the
substrate during the gait cycle, which is sufficiently large to detach
the particles. When the shoe is relatively far from the floor, the
induced airflow does not have a direct impact on the resuspension
process. This is because the airflow near the substrate is too low to
overcome the large adhesion force acting on the particles. However,
once the particles are detached from the surface and are suspended
in the air, the induced airflow plays an important role in the
dispersion and transport of particles in the environment.
Consistent with previous works [26,27], high speed airflows are
observed in the gap between the shoe and the flooring in both toe
and heel regions during the downward motion of the shoe. In
addition, the airflow field during the downward motion of the shoe
is much stronger in the toe region compared with that in the heel
section, which makes the toe part more effective for particle
resuspension. The generated velocity magnitude depends on the
average walking speed.
During upward motion, the shoe pivots at the tip of the shoe and
a strong suction flow is generated between shoe bottom and
108 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109

flooring that could resuspend additional particles. This motion and

the rotation of the shoe about the toe continue and lead to the
transport of generated vortices. These vortices are considered to be
the main mechanism of dispersion and transport of resuspended
particles to the environment.
Zhang et al. [27] presented a simplified model of a shoe with two
circular disks in toe and heel regions and predicted that there are
circular bands around the circles as the regions susceptible for
particle resuspension. Fig. 5 shows the variation of the shear ve-
locity on the flooring during the gait cycle. Similar to the model
predictions of Zhang et al. [27], a ring shape region near the edge of
the shoe is formed that experiences high values of shear velocity
that leads to more particle resuspension.

4.2. Comparison with experimental chamber study

Particle resuspension under the effects of shoe motion during

the gait cycle was evaluated using the resuspension apparatus ex- Fig. 8. Time variation of simulated number of 5e7.5 mm particles entering the sam-
periments, as well as the computational modeling approach. In the pling probe at the toe and center parts at the height of 5 cm during the gait cycle.
Comparison with the experimental data for mechanical shoe.
computer simulations the particle and flooring types and particle
loading identical to those of the experimental studies were used.
Table 2 provides the chemical composition of Arizona Test Dust simulation results are in close agreement with the experimental
(ATD) for which 68e75% is SiO2. Table 3 lists the material properties data, further improvement may be observed adding the
of ATD and vinyl flooring that are used for evaluation of resus- compressibility effect in the analysis [53].
pension shear velocity. The non-dimensional roughness parameter
Dc is varied in the range of 0.6e0.95 based on the earlier studies 5. Conclusions
[44,51,52]. In this work, the non-dimensional roughness parameter
is assumed have mean value of 0.82 based on the wind tunnel study In this study, particle resuspension due to human walking was
of Goldasteh et al. [49] using the same material for particle and investigated both numerically and experimentally. Experimental
flooring. Values of all parameters that are used in resuspension part of the study was conducted using a mechanical shoe resus-
simulation are listed in Table 4. pension apparatus. This setup was designed to mimic the human
Fig. 6 shows the sequence of resuspended particle distribution foot step during walking activity. In the numerical section, a RANS
during the first stepping cycle. This figure emphasis on the resus- approach with the RNG K-epsilon turbulence model was employed
pension of particles during downward motion (t  1) from a cir- for simulating the airflow field around and under the shoe during
cular region in toe section which is in good agreement with the the gait cycle. The model accounted for the particle adhesion to the
previous figures of Section 4.1. Also, during upward motion (t > 1) substrate and resuspension of particles by induced airflow gener-
the simulation shows the effects of suction airflow on pulling par- ated by movement of shoe during gait cycle. The resuspension
ticles away from the flooring and dispersing airborne particles. model of Goldasteh et al. [10,49], was included in the CFD package
Fig. 7 shows the sequence of the 3D state of particles during first through UDFs [37]. Tracking of resuspended particles were per-
8 s of simulation which clearly shows the resuspension and formed by a Lagrangian trajectory analysis approach. The time
dispersion of particles in gait cycle. Consistent with the experi- variation of number of particles that enter the sampling probes
mental observations, the motion of the shoe in gait cycle leads to located at the toe and side parts of the shoe were evaluated versus
the higher concentration of particles in front of the shoe rather than time. The simulated number of particles was shown to be in good
the side section. This figure shows that the motion of the shoe by agreement with experimental data for the mechanical shoe for
itself has an important effect on dispersion of the resuspended different size bins and various locations.
particles in the chamber. Simulations were performed for 100 s and
the number of particles entering the measuring probe inlets was
evaluated. Fig. 8 shows the variation of simulated number of par-
ticles versus time at the toe and center locations of probe inlets for
particles between 5 and 7.5 mm. This figure shows higher number
concentration of particles in the front of the shoe compared to the
side section during the gait cycle. Experimental results measured
for the mechanical foot are also reproduced in this figure for
comparison. It is seen that the simulation results are in close
agreement with the experimental data.
Fig. 9 shows the variation of simulated number of particles
entering the probe at the front section of the shoe versus time for
two different size bins of 5e7.5 mm and 7.5e10 mm particles. The
simulation results clearly show higher concentration of particle for
smaller size bin compared to the larger size, even though the larger
particles are easier to detach. This is primarily due to the particle
size distribution of ATD that have far more particles in the 5e7.5 mm
size range, which is also similar to that of house dust found in
homes [37]. The larger particles are also more strongly affected by Fig. 9. Variation of simulated particle number at the height of 5 cm at the toe part with
the gravity force resulting in higher deposition. Although the time. (5e7.5 mm vs. 7.5e10 mm).
 I. Goldasteh et al. / Building and Environment 77 (2014) 101e109
Close observation of shoe motion during the gait cycle showed
that the high speed airflows generated in the gap of the shoe sole
and flooring in both downward and upward motion are the main
mechanism for particle resuspension. In addition, a strong suction
flow generated during upward motion appears to be the main
mechanism for dispersion and transport of airborne particles in the
environment.
Comparison of the number concentrations shows that the shoe
motion generates higher rate of particle resuspension in front of the
shoe compared to the side section. Also, simulation results showed
the high number of resuspended particles in the size range of 5e
7.5 mm compared to 7.5e10 mm due to the size distribution of the
test particles.
Acknowledgments
This material is based upon work supported by the National
Science Foundation under Grant Number CBET 0846704. Any
opinions, findings, and conclusions or recommendations expressed
in this material are those of the authors and do not necessarily
reflect the views of the funding agency.
References
[1] Ferro AR, Kopperud RJ, Hildemann LM. Source strengths for indoor human
activities that resuspend particulate matter. Environ Sci Technol 2004;38:
1759e64.
[2] Robinson JP, Thomas J. Time spent in activities, locations, and microenviron-
ments: a California-national comparison. Project Report; 1991.
[3] Carrer P, Maroni M, Alcini D, Cavallo D. Allergens in indoor air: environmental
assessment and health effects. Sci Total Environ 2001;270:33e42.
[4] Kopperud RJ, Ferro AR, Hildemann LM. Outdoor versus indoor contributions to
indoor particulate matter (PM) determined by mass balance methods. J Air
Waste Manag Assoc 2004;54:1188e96.
[5] Thatcher TL, Layton DW. Deposition, resuspension, and penetration of parti-
cles within a residence. Atmos Environ 1995;29:1487e97.
[6] Nazaroff WW. Indoor particle dynamics. Indoor Air 2004;14:175e83.
[7] Abt E, Suh HH, Allen G, Koutrakis P. Characterization of indoor particle
sources: a study conducted in the metropolitan Boston area. Environ Health
Perspect 2000;108:35e44.
[8] Gillette DA, Lawson RE, Thompson RS. A “test of concept” comparison of
aerodynamic and mechanical resuspension mechanisms for particles depos-
ited on field rye grass (Secale cercele). Part 1. Relative particle flux rates. Atmos
Environ 2004;38:4789e97.
[9] Qian J, Tavakoli B, Goldasteh I, Ahmadi G, Ferro AR. Building removal of par-
ticulate pollutant plume during outdoor resuspension event. Build Environ
2014;75:161e9.
[10] Goldasteh I, Ahmadi G, Ferro A. A model for removal of compact, rough,
irregularly shaped particles from surfaces in turbulent flows. J Adhesion
2012;88:766e86.
[11] Biasi L, de los Reyes A, Reeks MW, de Santi GF. Use of a simple model for the
interpretation of experimental data on particle resuspension in turbulent
flows. J Aerosol Sci 2001;32:1175e200.
[12] Ibrahim AH, Dunn PF, Brach RM. Microparticle detachment from surfaces
exposed to turbulent air flow: controlled experiments and modeling. J Aerosol
Sci 2003;34:765e82.
[13] Ibrahim AH, Dunn PF, Brach RM. Microparticle detachment from surfaces
exposed to turbulent air flow: effects of flow and particle deposition char-
acteristics. J Aerosol Sci 2004;35:805e21.
[14] Ibrahim AH, Brach RM, Dunn PF. Microparticle detachment from surfaces
exposed to turbulent air flow: microparticle motion after detachment.
J Aerosol Sci 2004;35:1189e204.
[15] Goldasteh I, Ahmadi G, Ferro AR. Monte Carlo simulation of micron size
spherical particle removal and resuspension from substrate under fluid flows.
J Aerosol Sci 2013;66:62e71.
[16] Boor BE, Siegel JA, Novoselac A. Monolayer and multilayer particle deposits on
hard surfaces: literature review and implications for particle resuspension in
the indoor environment. Aerosol Sci Technol 2013;47:831e47.
[17] Wang S, Zhao B, Zhou B, Tan Z. An experimental study on short-time particle
resuspension from inner surfaces of straight ventilation ducts. Build Environ
2012;53:119e27.
[18] Kim Y, Gidwani A, Wyslouzil BE, Sohn CW. Source term models for fine par-
ticle resuspension from indoor surfaces. Build Environ 2010;45:1854e65.
[19] Zhu Y, Zhao B, Zhou B, Tan Z. A particle resuspension model in ventilation
ducts. Aerosol Sci Technol 2012;46:222e35.
109
[20] Ibrahim A, Dunn P, Qazi M. Experiments and validation of a model for
microparticle detachment from a surface by turbulent air flow. J Aerosol Sci
2008;39:645e56.
[21] Barth T, Reiche M, Banowski M, Oppermann M, Hampel U. Experimental
investigation of multilayer particle deposition and resuspension between
periodic steps in turbulent flows. J Aerosol Sci 2013;64:111e24.
[22] Goldasteh I, Ahmadi G, Ferro A. Effect of air flow on dust particles resus-
pension from common flooring. In: ASME FEDSM-ICNMM, 1; 2010. pp. 2797e
800.
[23] Qian J, Ferro AR. Resuspension of dust particles in a chamber and associated
environmental factors. Aerosol Sci Technol 2008;42:566e78.
[24] Madler L, Koch W. Particle resuspension from surfaces by impacting objects.
J Aerosol Sci 1997;28:85e6.
[25] Madler L, Koch W. Particle resuspension burst from dust layers induced by
impacting objects. J Aerosol Sci 1999;30.
[26] Khalifa HE, Elhadidi B. Particle levitation due to a uniformly descending flat
object. Aerosol Sci Technol 2007;41:33e42.
[27] Zhang X, Ahmadi G, Qian J, Ferro A. Particle detachment, resuspension and
transport due to human walking in indoor environments. J Adhesion Sci
Technol 2008;22(5):591e621.
[28] Kubota Y, Hall JW, Higuchi H. An experimental investigation of the flowfield
and dust resuspension due to idealized human walking. J Fluids Eng
2009;131:081104.
[29] Eisner A, Rosati J, Wiener R. Experimental and theoretical investigation of
particle-laden airflow under a prosthetic mechanical foot in motion. Build
Environ 2010;45:878e86.
[30] Oberoi RC, Choi JI, Edwards JR, Rosati JA, Thornburg J, Rodes CE. Human-
induced particle re-suspension in a room. Aerosol Sci Technol 2010;44:216e
29.
[31] Choi JI, Edwards J. Large-eddy simulation of human-induced contaminant
transport in room compartments. Indoor Air 2012;22:77e87.
[32] Choi JI, Edwards JR, Rosati JA, Eisner AD. Large eddy simulation of particle re-
suspension during a footstep. Aerosol Sci Technol 2012;46:767e80.
[33] Kubota Y, Higuchi H. Aerodynamic particle resuspension due to human foot
and model foot motions. Aerosol Sci Technol 2013;47:208e17.
[34] Tian Y, Sippola M, Sul K, Ferro AR. Development and validation of a consistent
experimental methodology to investigate walking-induced resuspension,
Aerosol Sci Technol; 2013 [submitted for publication].
[35] Tian Y, Sul K, Qian J, Mondal S, Ferro AR. A comparative study of walking-
induced dust resuspension using a consistent test mechanism. Indoor Air;
2014, [in press].
[36] Tian Y. Characterization of walking-induced particle resuspension using a
consistent test mechanism [PhD Thesis]. Potsdam, NY, USA: Clarkson Uni-
versity; 2013.
[37] Goldasteh I. Fundamentals and applications of particle resuspension, disper-
sion and transport in indoor environment [PhD Thesis]. Potsdam, NY, USA:
Clarkson University; 2013.
[38] Launder BE, Spalding D. The numerical computation of turbulent flows.
Comput Methods Appl Mech Eng 1974;3:269e89.
[39] Choudhury D. Introduction to the renormalization group method and turbu-
lence modeling. Fluent Inc; 1993. Technical Memorandum TM-107.
[40] Zhao B, Zhang Y, Li X, Yang X, Huang D. Comparison of indoor aerosol particle
concentration and deposition in different ventilated rooms by numerical
method. Build Environ 2004;39:1e8.
[41] Li A, Ahmadi G. Dispersion and deposition of spherical particles from point
sources in a turbulent channel flow. Aerosol Sci Technol 1992;16:209e26.
[42] Fuchs NA. The mechanics of aerosols. New York: Macmillan; 1964.
[43] Friedlander SK. Smoke, dust and haze: Fundamentals of aerosol behavior. New
York: Wiley; 1977.
[44] Soltani M, Ahmadi G. In: Detachment of rough particles with electrostatic
attraction from surfaces in turbulent flows, 13; 1999. pp. 325e55.
[45] Goldasteh I, Ahmadi G, Ferro AR. Monte Carlo simulations of micro-particle
detachment and resuspension from surfaces in turbulent flows. In: ASME
FEDSM, 1; 2012. pp. 259e65.
[46] Fuller KNG, Tabor D. The effect of surface roughness on the adhesion of elastic
solids. Proc R Soc Lond 1975;345:327e42.
[47] Rabinovich YI, Adler JJ, Ata A, Singh RK, Moudgil BM. Adhesion between
nanoscale rough surfaces: II. Measurement and comparison with theory.
J Colloid Interface Sci 2000;232:17e24.
[48] Rabinovich YI, Adler JJ, Ata A, Singh RK, Moudgil BM. Adhesion between
nanoscale rough surfaces: I. Role of asperity geometry. J Colloid Interface Sci
2000;232:10e6.
[49] Goldasteh I, Ahmadi G, Ferro AR. Wind tunnel study and numerical simulation
of dust particle resuspension from indoor surfaces in turbulent flows. J Adhes
Sci Technol; 2012:1e17.
[50] O’Neill ME. A sphere in contact with a plane wall in a slow linear shear flow.
Chem Eng Sci 1968;23:1293e8.
[51] Soltani M, Ahmadi G. On particle adhesion and removal mechanisms in tur-
bulent flows. J Adhes Sci Technol 1994;8:763e85.
[52] Soltani M, Ahmadi G. Particle removal mechanisms under substrate acceler-
ation. J Adhesion 1994;44:161e75.
[53] Elhadidi B, Khalifa HE. Aerodynamic resuspension of particles due to a falling
flat disk. Part Sci Technol 2013;31:35e44.