ME 437/ME 537

PARTICLE TRANSPORT, DEPOSITION

AND REMOVAL

Goodarz Ahmadi

Department of Mechanical
and Aeronautical Engineering
Clarkson University
Potsdam, NY 13699-5725

1
 Air pollution and smog.

Particle trajectories in a hot gas filtration vessel.

2
Sample glass fiber particles.

3
Dust storm over the red sea.

4
Birth place of stars.

5
 INTRODUCTION TO AEROSOLS

Definition: Aerosol is a suspension of solid or liquid particles in a gas. Dust, smoke,

mists, fog, haze, and smog are various forms of common aerosols. Aerosol particles are
found in different shapes (isometrics, platelets, and fibers) and different sizes. For
irregular shaped particles, different equivalent diameters are defined. Examples of
equivalent diameters are:

• Equivalent area diameter,

• Feret’s diameter (maximum distance edge to edge);
• Stoke’s diameter (diameter of a sphere with the same density and the same velocity as
the particle);
• Aerodynamic diameter (diameter of a sphere with the density of water and the same
velocity as the particle).

The range of diameters of common aerosol particles is between 0.01 and 100 µm.
The lower limit of 10 nm roughly corresponds to the transition from molecule to particle.
Particles larger than 100 µm normally do not remain suspended in air for a sufficient
amount of time. Noting that the mean free path for air is about 0.07 µm and visible light
has a wavelength band of 0.4 – 0.7 µm, the mechanical and optical behaviors of particles
are significantly affected by their size.

Particles greater than 5 – 10 µm are usually removed by the upper respiratory

system. But particles smaller than 5 µm can penetrate deep into the lung and become a
health hazard. Typical ranges of values for aerosol parameters for aerosols are listed in
Table 1. The corresponding values for air (N 2 ) are also shown in this table for
comparison.

Table 1 - Parameters of Aerosol in the Atmosphere

Aerosols Air
Number Density 100-10 5
10 19
(Number/cm 3 )
Mean Temperature (K) 240 – 310 240 – 310
Mean Free Path Greater than 1m 0.06 µm
Particle Radius 0.01 – 10 µm 2 × 10 −4 µm
Particle Mass (g) 10 −18 - 10 −9 4.6 × 10 −23
Particle Charge (in 0 – 100 Weakly Ionized Single
Elementary Charge Units) Charge

The important relevant dimensionless groups relevant the motion of aerosols are
listed in Table 2.

6
 Table 2 – Dimensionless Groups
Knudsen Number 2λ
Kn =
d
Mach Number | vp − vf |
M=
cf
Schmidt Number ν n f λd 2
Sc = =
D 4
Brown Number v p,2 | v' p |
Br = ( )1 / 2 =
v f ,2 | v'f |
Reynolds Number | v p − v f | d 4M
Re = =
ν Kn

Here the following symbols are defined:

λ = Mean Free Path ν = Kinematic Viscosity

d = Particle Diameter D = Diffusivity
v p = Particle Velocity v’ = Thermal Velocity
v f = Fluid (Air) Velocity n = Number Density
c f = Speed of Sound

Here superscript " f " corresponds to fluid and superscript " p " denotes particle.

In these equations the root mean square fluctuation velocity is given by

| v'f | = (8kT / πm f )1 / 2
and

ν = 0.5 c f λ

The mean free path of the gas is given as

1 kT
λ= =
2πnd 2πd 2m P
2
m

Here n is the gas number density, d m is the gas molecule (collisional) diameter,
k = 1.38 × 10 -23 J/K is the Boltzmann constant, P is pressure, and T is temperature. For
air, d m = 0.361 nm and
23.1T
λ(µm) = , P is in Pa, and T is K.
P

7
 Table 3. Aerosol Characteristics

Particle Diameter, µm
−4 −3 −2
10 10 10 10 −1 10 0 101 10 2 10 3 10 4

Electromagnetic x-Ray UV Vis Infrared Microwaves

Wave

Definition Solid Fume Dust

Liquid Mist Spray

Soil Clay Silt Sand Gravel

Atmospheric Smog Cloud/Fog Mist Rain

Typical Viruses Bacteria Human Hair

Particles Smoke Coal Dust Beach Sand

Size Analysis Microscope

Method Electron Microscope Sieving
x-Ray Diffraction
Ultra Centrifuge Sedimentation

Gas Cleaning Ultrasonics Settling Chamber

Centrifugal
Liquid Scrubber
Air Filter
HE Air Filter Impact Separators
Thermal Separators
Electrostatic Separators

Diffusion Air 5 × 10 −2 5 × 10 −4 10 −5 3 × 10 −7 2 × 10 −9 2 × 10 −10 2 × 10 −11

Coeff. cm 2 / s Water 5 × 10 −6 5 × 10 −7 5 × 10 −8 5 × 10 −9 5 × 10 −10 5 × 10 −11 5 × 10 −12 5 × 10 −13

Terminal (S=2) Air 10 −6 1.5 × 10 −5 2 × 10 −4 7 × 10 −3 0.6 50 600 2.5 × 10 3

Velocity cm / s Water 10 −10 6 × 10 −9 6 × 10 −7 6 × 10 −5 6 × 10 −3 0 .6 12 58

10 −4 10 −3 10 −2 10 −1 10 0 101 10 2 10 3 10 4

Particle Diameter, µm

8
HYDRODYNAMIC FORCES

Drag Force and Drag Coefficient

A particle suspended in a fluid is subjected to hydrodynamic forces. For low

Reynolds’ number, the Stokes drag force on a spherical particle is given by

F D = 3πµUd, (1)

where d the particle diameter, µ is the coefficient of viscosity and U is the relative
velocity of the fluid with respect to the particle. Equation (1) may be restated as

FD 24
CD = = (2)
1
ρU 2 A Re
2

πd 2
In Equation (3), ρ is the fluid (air) density, A = is cross sectional area of the
4
spherical particle, and

ρUd
Re = (3)
µ

is the Reynolds number. The Stokes drag is applicable to the creeping flow regime
(Stokes regime) with small Reynolds numbers (Re < 0.5). At higher Reynolds numbers,
the flow the drag coefficient deviates from Equation 2. Figure 1 shows the variation of
drag coefficient for a sphere for a range of Reynolds numbers.

Eq. (4)
CD
Eq. (2)

Re
Figure 1. Variations of drag coefficient with Reynolds number for a spherical particle.

9
 Oseen included the inertial effect approximately and developed a correction to the
Stokes drag given as

24[1 + 3 Re/ 16]

CD = , (4)
Re

which is shown in Figure 1.

For 1 < Re < 1000, which is referred to as the transition regime, the following
expressions may be used (Clift et al., 1978):

24[1 + 0.15 Re 0.687 ]

CD = , (5)
Re

or

24 4
CD = + 0.33 (6)
Re Re

1000

100
Experiment
CD

10 Oseen

Stokes Eq. (5)

1 Newton

0
0 1 10 100 1000 10000
Re
Figure 2. Predictions of various models for drag coefficient for a spherical particle.

10
 For 10 3 < Re < 2.5 × 10 5 , the drag coefficient is roughly constant ( C D = 0.4 ). This
regime is referred to as the Newton regime. At Re ≈ 2.5 × 10 5 , the drag coefficient
decreases sharply due to the transient from laminar to turbulent boundary layer around
the sphere. That causes the separation point to shift downstream as shown in Figure 3.

Laminar Boundary Layer Turbulent Boundary Layer

Figure 3. Laminar and turbulent boundary layer separation.

Wall Effects on Drag Coefficient

For a particle moving near a wall, the drag force varies with distance of the
particle from the surface. Brenner (1961) analyzed the drag acting on a particle moving
toward a wall under the creeping flow condition as shown in Figure 4a. To the first
order, the drag coefficient is given as

24 d
CD = (1 + ) (7)
Re 2h

d U
d

h h
U

(a) Motion normal to the wall (b) Motion parallel to the wall

Figure 4. Particle motions near a wall.

For a particle moving parallel to the wall as shown in Figure 4b, the Stokes drag
force need to be modifies. For large distances from the wall, Faxon (1923) found

24 9 d 1 d 45 d 4 1 d 5 −1
CD = [1 − ( ) + ( ) 3 − ( ) − ( ) ] (8)
Re 16 2h 8 2h 256 2h 16 2h

11
Cunningham Correction Factor

For very small particles, when the particle size becomes comparable with the gas
mean free path, slip occurs and the expression for drag must be modified accordingly.
Cunningham obtained the needed correction to the Stokes drag force:

3πµUd
FD = , (9)
Cc

where the Cunningham correction factor C c is given by

2λ
Cc = 1 + [1.257 + 0.4e −1.1d / 2 λ ] (10)
d
Here λ denotes the molecular mean free path in the gas. Note that C c > 1 for all values
of d and λ. Figure 5 shows the variation of Cunningham correction factor with Knudsen
number. It is seen that C c is about 1 for Kn <0.1 and increases sharply as Kn increases
beyond 0.5. Table 4 illustrates the variation of Cunningham correction factor with
particle diameter in air under normal pressure and temperature conditions with λ = 0.07
λ
µm. Equation (10) is applicable to a wide range of Kn = ≤ 1000 that covers slip,
d
transition and part of free molecular flows. The particle Reynolds number and Mach
number (bases on relative velocity), however should be small.

1000

100
Cc

10

1
0.001 0.01 0.1 1 10 100
Kn

Figure 5. Variation of Cunningham correction with Knudsen number.

12
 Table 4 – Variations of C c with d for λ = 0.07 µm
Diameter, µm Cc
10 µm 1.018
1 µm 1.176
0.1 µm 3.015
0.01 µm 23.775
0.001 µm 232.54

Compressibility Effect

For high-speed flows with high Mach number, the compressibility could affect the

drag coefficient. Many expressions were suggested in the literature to account for the

effect of gas Mach number on the drag force. Henderson (1976) suggested two

expressions for drag force acting on spherical particles for subsonic and supersonic flows.

Accordingly, for subsonic flow

−1
   Re
CD = 24Re+ S 4.33 + 1.567× exp − 0.247 
   S 

+ exp − 
(
 0.5M  4.5 + 0.38 0.03Re+ 0.48 Re )    M 
+ 0.1M 2 + 0.2M 8  + 1 − exp − 0.6S
 Re  1 + 0.03Re+ 0.48 Re    Re

(11)

where M is Mach number based on relative velocity, ∆V = V − V p , and S= M γ is

2

the molecular speed ratio, where γ is the specific heat ratio. For the supersonic flows

with Mach numbers equal to or exceeding 1.75, the drag force is given by

1
0.34 M 2  2 1.058 1 
0.9 + 2 + 1.86  2 + 2 + − 4
M  Re   S S S 
CD = 1
(12)
M  2
1 + 1.86 
 Re 

For the flow regimes with Mach between 1 and 1.75, a linear interpolation is to be used.

13
 Carlson and Hoglund (1964) proposed the following expression:

0.427 3
1 + exp(− − 0.88 }
24 M 4.63
Re
CD = (13)
Re M Re
1+ {3.82 + 1.28 exp(−1.25 )}
Re M

Droplets

For drag force for liquid droplets at small Reynolds numbers is given as

1 + 2µ f / 3µ p
FD = 3πµ f Ud (14)
1 + µf / µp

where the superscripts f and p refer to the continuous fluid and discrete particles
(droplets, bubbles), respectively.

Non-spherical Particles

For non-spherical (chains or fibers) particles, Stokes’ drag law must be modified.
i.e.,

FD = 3πµUd e K , (15)

where d e is the diameter of a sphere having the same volume as the chain or fiber. That
is,

6
d e = ( Volume)1 / 3 (16)
π

and K is a correction factor.

For a cluster of n spheres, d e = n 1 / 3 d . For tightly packed clusters, k < 1.25.

Some other values of K are listed in Table 5.

14
 Table 5 – Correction Coefficient
Cluster Correction Cluster Correction Cluster Correction
Shape Shape Shape
oo K = 1.12 oooo K = 1.32 oo K = 1.17
oo
ooo K = 1.27 ooooo K = 1.45 oo K = 1.19
o
oo
o K = 1.16 oooooo K = 1.57 oo K = 1.17
oo oo
oo
oooooo K = 1.64 ooooooo K = 1.73
o
o

Ellipsoidal Particles

For particles that are ellipsoids of revolution, the drag force is given by

FD = 6πµUaK' (17)

where a is the equatorial semi-axis of the ellipsoids and K’ is a shape factor.

For the motion of a prolate ellipsoid along the polar axis as shown in Figure 6a,

b b

a a

(a) (b)

Figure 6. Motions of prolate ellipsoids in a viscous fluid.

15
 4 2
(β − 1)
3 b
K' = (β = ) (18)
( 2β − 1)
2
a
ln[β + (β 2 − 1)1 / 2 ] − β
(β 2 − 1)1 / 2

where β is the ratio of the major axis b to the minor axis a.

For the motion of a prolate ellipsoid of revolution transverse to the polar axis, as
shown in Figure 6b

8 2
(β − 1)
3 b
K' = (β = ) (19)
( 2β − 3)
2
a
ln[β + (β 2 − 1)1 / 2 ] + β
(β − 1)
2 1/ 2

Similarly for the motion of an oblate ellipsoid of revolution along the polar axis as
shown in Figure 7a,

4 2
(β − 1)
3 a
K' = (β = ) (20)
β(β − 2)
2
b
tan −1 (β 2 − 1)1 / 2 ] + β
(β − 1)
2 1/ 2

b b

a a

(a) (b)

Figure 7. Motions of oblate ellipsoids in a viscous fluid.

For the motion of an oblate ellipsoid transverse to the polar axis as shown in Figure 7b,

16
 8 2
(β − 1)
3 a
K' = (β = ) (21)
β(3β − 2)
2
b
tan −1 (β 2 − 1)1 / 2 ] − β
(β 2 − 1)1 / 2

By taking the limit as β → ∞ in Equations (17)-(21), the drag force on thin disks
and needles may be obtained. These are:

Thin Disks of Radius a

For motions perpendicular to the plane of the disk as shown in Figure 8a

FD = 16µaU (22)

For motions along the plane of the disk as shown in Figure 8b

FD = 32µaU / 3 (23)

a a

(a) (b)

Figure 8. Motions of a thin disk in a viscous fluid.

Ellipsoidal Needle of Length 2b

For motions along the needle as shown in Figure 9a

4πµUb b
FD = , (β = ) (24)
ln 2β a

For side way motions of the needle as shown in Figure 9b

8πµUb
FD = (25)
ln 2β

17
 b b

(a) (b)

Figure 9. Motions of a needle in a viscous fluid.

Cylindrical Needle

For a cylindrical needle with a very large ratio of length to radius ratio, moving
transverse to its axis as shown in Figure 10, the drag per unit length is given as

4πµU
FD = (26)
(2.002 − ln R e )

2aU
where R e = and a is the radius. It is understood that
ν

Figure 10. Flow around a cylindrical needle.

18
Particle Shape Factor

The ratio of the resistance of a given particle to that of a spherical particle having
the same volume is called the dynamic shape factor of the particle, K. The radius of an
equal volume sphere is referred to as the equivalent radius r e . Clearly

re = αβ1/ 3 for prolate spheroids, (27)

re = αβ −1 / 3 for oblate spheroids. (28)

Hence,
K = K ' β1/ 3 for prolate ellipsoids, (29)

K = K ' β −1/ 3 for oblate ellipsoids. (30)

The Stokes (sedimentation radius) of a particle is the radius of a sphere with the same
density, which is settling with the terminal velocity of the particle in a quiescent fluid.
Values of shaped factors for a number of particles are available (Hidy, 1984; Lerman,
1979).

19
AEROSOL PARTICLE MOTION

Equation of Motion

Consider an aerosol particle in fluid flow as shown in Figure 1. The equation of

motion of a spherical aerosol particle of mass m and diameter d is given as

du p 3πµd f
m = (u − u p ) + mg (1)
dt Cc

Here u p is the particle velocity, u f is the fluid velocity, g is the acceleration of gravity
and the buoyancy effect in air is neglected. Here it is assume that the particle is away
from walls and the Stokes drag is assumed.

Drag

Gravity

Figure 1. Schematics of an aerosol motion in a gas flow.

3πµd
Dividing Equation (1) by and rearranging, we find
Cc
du p
τ = (u f − u p ) + τg (2)
dt
where the particle response (relaxation) time is defined as

mC c d 2 ρ p C c Sd 2 C c
τ= = = ,
3πµd 18µ 18ν
(3)

πd 3ρ p
where m = , ν is the kinematic viscosity of the fluid and S = ρ p / ρ f is the density
6
ratio. In practice, for non-Brownian particles, Cc ≈ 1 and

20
 d 2ρ p
τ≈ (4)
18µ

Terminal Velocity

For a particle starting from rest, the solution to (2) is given as

u p = (u f + τg)(1 − e − t / τ ) (5)

where u f is assumed to be a constant vector. For u f = 0 and large t, the terminal

velocity of particle u t is given by

ρ p d 2 gC c
u t = τg = (6)
18µ

Table 7 – Relaxation time τ for a unit density particle in air (p = 1 atm, T = 20 o C).
Diameter, µm ut = τ g τ sec Stop Distance Stop Distance
u o = 1 m/s u o = 10 m/s
0.05 0.39 µm/s 4 × 10 −8
0.04 µm 4 × 10 −4 mm
0.1 0.93 µm/s 9.15 × 10 −8 0.092 µm 9.15 × 10 −4 mm
0.5 10.1 µm/s 1.03 × 10 −6 1.03 µm 0.0103 mm
1 35 µm/s 3.57 × 10 −6 3.6 µm 0.0357 mm
5 0.77 mm/s 7.86 × 10 −5 78.6 µm 0.786 mm
10 3.03 mm/s 3.09 × 10 −4 309 µm 3.09 mm
50 7.47 cm/s 7.62 × 10 −3 7.62 mm 76.2 mm

Stopping Distance

In the absence of gravity and fluid flow, for a particle with an initial velocity of
p
u , the solution to (2) is given by
o

x p = u po τ(1 − e − t / τ ) (7)

u p = uoe − t / τ (8)

where x p is the position of the particle. As t → ∞, u p → 0 and

x p = u op τ (9)

21
is known as the stopping distance of the particle. For an initial velocity of 1000 cm/s, the
stop distance for various particles are listed in table 7.

Particle Path

For constant fluid velocity, integrating Equation (5), the position of the particle is
given by

x p = x po + u po τ(1 − e − t / τ ) + (u f + τg)[t − τ(1 − e − t / τ )] (11)

Here x op is the initial position of the particle. For a particle starting from rest, when the
fluid velocity is in x-direction and gravity is in the negative y-direction, Equation (10)
reduces to

x p / u f τ = [ t / τ − (1 − e − t / τ )] (12)

y p / u f τ = −α[ t / τ − (1 − e − t / τ )] (13)

where the ratio of the terminal velocity to the fluid velocity α is given by

τg
α= (14)
uf τ

Figure 2 shows the variation of vertical position of the particle with time.

0
α =0.1
-2

-4
α =1
y/utau

-6

-8
α =2
-10

-12
0 1 2 3 4 5 6
t/tau
Figure 2. Variations of the particle vertical position with time.

22
 From Equations (12) and (13), it follows that

y p = − αx p (15)

That is the particle paths are straight lines. Figure 3 shows sample particle trajectories.

0
α =0.1
-2

-4
y/utau

α =1
-6

-8

-10
α =2
-12
0 1 2 3 4 5 6
x/utau

Figure 3. Sample particle trajectories.

Buoyancy Effects

For small particles in liquids, the buoyancy effect must be included. Thus,
Equation (1) is replaced by

du p 3πµd f
(m + m a ) = (u − u p ) + (m − m f )g (16)
dt Cc

where m f is the mass of the equivalent volume fluid given as

πd 3ρ f
mf = (17)
6

and m a is the apparent mass with ρ f being the fluid density. For spherical particles,
1
ma = mf .
2

23
 Keeping the same definition for particle relaxation time as given by (3), Equation
(2) may be restated as

1 du p 1
(1 + )τ = (u f − u p ) + τg(1 − ) (18)
2S dt S

The expression for the terminal velocity then becomes

1 ρ p d 2 gC c ρf
u t = τ g (1 − ) = (1 − p ) (18)
S 18µ ρ

Note that the Basset force and the memory effects are neglected in this analysis.

24
Lift Force

Small particles in a shear field as shown in Figure 2 experience a lift force

perpendicular to the direction of flow. The shear lift originates from the inertia effects in
the viscous flow around the particle and is fundamentally different from aerodynamic lift
force. The expression for the inertia shear lift was first obtained by Saffman (1965,
1968). That is

du f 1 / 2 du f
FL (Saff ) = 1.615ρν1 / 2 d 2 (u f − u p ) | | sgn( ) (1)
dy dy

Here u f is the fluid velocity at the location of mass center of the particle, u p is the
du f
particle velocity, γ& = is the shear rate, d is the particle diameter and ρ and ν are the
dt
fluid density and viscosity. Note that F L is in the positive y-direction if u f >u p .

Lift

uf up

Figure 1. Schematics of a particle in a shear flow.

Equation (1) is subjected to the following constraints:

| uf − up | d Ωd 2
R es = << 1 R eΩ = << 1
ν ν
γ& d 2 R 1eG/ 2
R eG = << 1 ε= >> 1
ν R es

Here Ω is the rotational speed of the sphere. Dandy & Dwyer (1990) found that the
Saffman lift force is approximately valid at larger R es and small ε . McLaughlin (1991)
showed that the lift force decreases as ε decreases. Based on these studies Mei (1992)
suggested the following empirical fit to the results of Dandy and Dwyer and McLaughlin.
For large ε and R es ,

25
 FL (1 − 0.3314α1 / 2 ) exp(−R es / 10) + 0.3314α1 / 2 for R es ≤ 40
=
FL (Saff )  0.0524(αR es )1 / 2 for R es > 40
(2)

where
γ&d R es ε 2 R eG
α= = = (3)
2 | uf − up | 2 2R es

For 0.1 ≤ ε ≤ 20

FL
= 0.3{1 + tanh[2.5 log10 (ε + 0.191)]}{0.667 + tan[6(ε − 0.32)]} (4)
FL (Saff )

For large and small ε McLaughlin obtained the following expressions

FL  1 − 0.287ε −2 for ε >> 1

= −2
(5)
− 140ε ln(ε ) for ε << 1
5
FL (Saff )

Note the change in sign of the lift force for small values of ε .

McLaughlin (1993) included the effects of presence of the wall in his analysis of
the lift force. The results for particles in a shear field but not too close to the wall were
given in tabulated forms. Cherukat and McLaughlin (1994) analyzed the lift force acting
on spherical particles near a wall as shown in Figure 2. Accordingly

FL ( C − L ) = ρV 2 d 2 I L / 4 (6)

where

V = u p − u f = u p − γ& l

and for non-rotating spheres,

I L = (1.7716 + 0.216K − 0.7292K 2 + 0.4854K 3 )

− (3.2397 / K + 1.145 + 2.084K − 0.9059K 2 ) ∧ G (7)
+ (2.0069 + 1.0575 − 2.4007K 2 + 1.3174K 3 ) ∧ G2

26
 Lift

d l

Figure 2. Schematics of a particle near a wall in a shear flow.

For rotating (freely) spheres,

I L = (1.7631s + 0.3561K − 1.1837K 2 + 0.845163K 2 )

− (3.24139 / K + 2.6760 + 0.8248K − 0.4616K 2 ) ∧ G (8)
+ (1.8081 + 0.879585K − 1.9009K + 0.98149K ) ∧ 2 3 2
G

Here

d γ& d
K= , ∧G = . (9)
2l 2v

Lift Force on a Particle Touching a Plane

Leighton and Acrivos (1985) obtain the expression for the lift on the spherical
particles resting on a plane substrate as shown in Figure 4. They found

FL ( L −A ) = 0.576ρd 4 γ& 2 (10)

which is always point away from the wall. Note that the Saffman expression given by (1)
may be restated as

FL (Saff ) = 0.807ρν1 / 2 d 3 γ& 3 / 2 (11)

27
 Lift

Figure 3. Schematics of a sphere resting on a wall in a shear flow.

Equation (6) with I L given by (7) reduces to (10) for K = 1, ٨ G = -1.

For small particles in turbulent flows, using

+ + u + +yu * u *2
u =y , u = * , y = , γ= (12)
u ν ν

where u * is the shear velocity, equations (10) and (11) become

FL+( L − A ) = 0.576d +4 (13)

+ +3
FL (Saff ) = 0.807d (14)
where
FL du *
FL+ = , d+ = (15)
ρν 2 ν

Experimental studies of lift force were performed for generally larger particles in the
range of 100 to several hundred µm. Hall (1988) found

FL+( Hall) = 4.21d +2.31 for d + > 1.5 (16)

Mollinger and Nieuwstadt (1996) found

FL+( MN ) = 15.57d +1.87 for 0.15 < d + < 1 (17)

Figure 4 compares the model predictions with the experimental data of Hall. It is seen
that the experimental data is generally much higher than the theoretical models.

28
 1.00E+03
1.00E+02 Experiment

1.00E+01 Saffman
Mollinger
1.00E+00
Fl+

1.00E-01
Hall
1.00E-02
1.00E-03
Leighton
1.00E-04
1.00E-05
0.1 1 10
d+

Figure 4. Comparison of model predictions with the experimental data.

29