simulate the drying process of porous particle in a pneumatic
conveying dryer. The model takes into account the momentum, heat
and mass transfer between the continuous phase and the dispersed
phase. A single particle model was employed to calculate the
evaporation rate. In this model the pore structure is simplified to
allow the dominant evaporation mechanism to be readily identified at
all points within the duct. The predominant mechanism at any time
depends upon the pressure, temperature and the diameter of pore
from which evaporating is occurring. The model was validated
against experimental studies of pneumatic transport at low and high
speeds as well as pneumatic drying. The effects of operating
conditions on the dryer parameters are studied numerically. The
present results show that the drying rate is enhanced as the inlet gas
temperature and the gas flow rate increase and as the solid mass flow
rate deceases. The present results also demonstrate the necessity of
measuring the inlet gas velocity or the solid concentration in any
experimental analysis.
Keywords—Twophase, gassolid, pneumatic drying, pneumatic
conveying, heat and mass transfer
I. INTRODUCTION
RYING is an essential operation in the chemical, food,
agricultural, ceramic, polymers and plastic, pulp and
paper, pharmaceutical and wood processing industries.
One of the most widely used drying systems is the
pneumatic dryer and is also known as flash dryer, which can
be characterized as continuousconvective dryer. Pneumatic
dryers are characterized by simultaneous momentum, heat and
mass transfer processes between the dispersed material and
the drying agent. The large surface area for heat and mass
transfer result in higher drying rate and higher drying
capacity. In these types of dryers the contact time between the
drying medium and particulate material is relatively short
(usually few seconds only). Therefore, these dryers are
suitable for heatsensitive materials and also for removing
external moisture.
This allows higher inlet temperatures to be used than in
many other dryers without unduly heating the product [1].
Pneumatic dryers are simple in construction and have low
1
Faculty of Engineering, Menoufiya University, Shebin Elkom, Egypt,
s_elbehery@yahoo.com
2
Faculty of Engineering, Kafrelsheikh University, Kafrelsheikh, Egypt,
mofrehhh@yahoo.com
capital cost. Vertical type of construction, which facilitates
installation in exiting buildings, is an advantage of pneumatic
dryer systems [2].
The short residence time can, however, be a problem as an
advantage. Small fluctuation in the feed rate can suddenly
alter the heat and the mass balance along the tube and gives
very different exit conditions. In order to solve the complexity
of the process, many articles related to the theoretical and
mathematical model have been published.
By employing a volumetric heat transfer concept, as used
for rotary dryers, simple estimation procedures have been
suggested by Perry and Chilton [3]. These procedures
assumed that the particles were traveling at a steady velocity
close to the gas velocity. Baeyens et al [1] pointed out that
these methods can overpredict the required dryer length by
200% to 400%.
To model the acceleration zone accurately, a stepwise
procedure has been suggested by many workers including
Thorpe et al. [4], Kemp et al. [5] and Kemp and Oakley [6].
Although these procedures are considerable improvement on
the steadystate, Kemp et al. [7] reported that they can still
give errors of 50100% in the tube length prediction. Baeyens
et al. [1] and Radford [8] neglected the effect of acceleration
zone near the feed point in their stepwise procedure.
A steady state onedimensional model for pneumatic
drying of wet particle was presented by Levy and Borde [9].
They assumed a twostage drying process, with mass transfer
controlled by evaporation from a saturated outer particle
surface in the first stage and by diffusion within the particle in
the second stage. The model predictions were compared with
the experimental data obtained in large scale and pilot scale
pneumatic dryers and a good agreement was obtained.
Skuratovsky et al. [1011] developed a twodimensional
steadystate model based on the twofluid theory. The
predictions of the model were compared to the same
experimental data used by Levy and Borde [9]. The
predictions of the twodimensional model did not present any
significant difference as compared to those provided by Levy
and Borde [9]. However, the radial distribution of gas and
solid velocities and moisture content results in uneven cross
sectional, which may be an important aspect to be considered
in drying processes.
Pelegrina and Crapiste [12] presented a onedimensional
model for drying of food particles. The model took into
account the particle shrinkage during the drying process and
the non spherical shape of the particle was considered in drag
Porous Particles Drying in a Vertical Upward
Pneumatic Conveying Dryer
Samy M. ElBehery
1
, W. A. ElAskary
1,
K. A. Ibrahim
1
and Mofreh H. Hamed
2
D
International Journal of Aerospace and Mechanical Engineering 5:2 2011
110
and heat transfer coefficients. They assumed that the internal
resistance does not control the mass and energy transfer
between solid particles and air. They found that, in the low
range of air flow rates; the pressure drop under drying
conditions is higher than that under transport conditions. An
opposite effect was observed at higher velocities. However,
the model was not verified with experimental results.
Narimatsu et al. [13] investigated numerically and
experimentally the drying process of porous alumina and solid
glass particles in a vertical dryer. The model was for one
dimensional incompressible flow and the internal resistance
did not control the heat and mass transfer. Dry solids were
used in heat transfer experiments, and the measurements of
heat transfer coefficient indicted that the maximum value of
heat transfer coefficient occurred at the velocity of minimum
pressure drop. Furthermore, it was noticed that the
morphology of particles (porous or non porous) did not
influence in the air temperature profiles.
Fyhr and Rasmuson [1415] presented a more complex
model for a pneumatic dryer considering a distribution of
particle sizes for steam drying of wood chips. The model
includes a comprehensive twodimensional model for single
particle drying of single wood ship and onedimensional plug
flow was assumed. The irregular movement and the non
spherical shape of the wood ships were accounted by
measuring drag and heat transfer coefficients. To validate the
model, measurements of the temperature and pressure profiles
as well as the final moisture content were carried out, and the
predictions agreed well with the experimental results.
Unlike the above studies, which were performed in a
vertical upward pneumatic dryer, Alvarez et al. [16] have
studied numerically and experimentally the drying process in
a vertical downward pneumatic dryer. The model was for non
shrinkage spherical particle and steady state onedimensional
flow. Some experimental works on the pneumatic dryer were
given also,by [1719].
The present paper concerns with a onedimensional model
for a pneumatic drying of porous particles. The model
formulations are similar to that of Levy and Borde [9] but the
present model allows for higher temperature. The mass
transfer process was obtained by the comprehensive single
particle model described by Radford [8]. The model of Levy
and Borde [9] assumes that the drying process stops and
particle breakup occurs if the saturation pressure inside the
wet core is greater than the ambient pressure. The gas pressure
in the pneumatic drying process is around atmospheric
pressure, therefore, according to their assumption, particle
breakup may be occurred if the temperature inside the wet
core exceeds 100oC at any point within the duct. Also, it is
important to notice that the difference between the present
model and that of Radford [8] is that Radford's model assumes
constant pressure along the duct and neglects the velocity
difference between the two phases.
II. MATHEMATICAL MODELING
To formulate the suggested model, a quasione dimensional
situation has been considered. The model is concerned with
the twophase flow of gas and particles through a vertical
pipe. The model which was developed by Hamed [20] has
been modified and used here based on the following
assumptions:
 The flow is onedimensional and steady.
 The particles are spherical in shape.
 The gas phase is mixture of water vapor and dry
air.
 All particles interaction is ignored in the model.
This implies that any mass, heat or momentum
transfer between particles is insignificant to the
transfer between individual particles and the gas
stream.
 The model assumes that the solid will be conveyed
as discrete particles and that heat and mass transfer
occur from individual particles.
A. Governing Equations
Based on the above mentioned assumptions the governing
equations for the gas and dispersed phases are derived
according to the basic laws of fluid mechanics as follows:
 The mass balance equation for the gas phase may be
written as:
( )
g g g mass
d
u A S
dx
o p = (1)
 The momentum equation for the gas phase can be
expressed as:
2
( )
g g g g g
wg mom mass d
d dP
u A A gA
dx dx
F S S u
o p o p = ÷ ÷
÷ + +
(2)
 The total energy equation for the gas phase can be written
as:
2
2
2
2
g
g g g g wall
d
g g g mass wv energy
u
d
u A H Q
dx
u
u Ag S H S
o p
o p
 
+ = 

\ .
 
÷ + + +

\ .
(3)
Where, S
mass
, S
mom
and S
energy
are mass, momentum
and energy coupling source terms, respectively.
 The equation of motion for a particle in a gas is given by:
( )
2
3
2
2 1
g D
d
g d g d
d p
g d d
p
d
C
du
u u u u
dx d
u u
g f
d
p
p
p
p
= ÷ ÷
 
÷ ÷ ÷

\ .
(4)
International Journal of Aerospace and Mechanical Engineering 5:2 2011
111
where, C
D
is the drag coefficient and it is calculated as given
in [21] as:
5
646 . 0
10 3 Re 400 5 . 0
400 Re 1
Re
24
1 Re
Re
24
× < < =
s < =
s =
p
p
p
p
p
D
C
where, Re /
p g p g d g
d u u p u = ÷ is the particle Reynolds
number.
The equation for particle temperature, assuming the
temperature is uniform throughout the particle can be written
as:
2
( )
d
d p pd p g d
o
d
fg
dT
u m C d h T T
dx
m H
;t = ÷
÷
(5)
The residence time of the particles at the gas phase was
calculated as given in [9] as:
d
d
u dx
dt 1
= (6)
B. Coupling between phases
An important concept in the analysis of twophase gassolid
flows is to consider the mutual effect between the two phases.
The mass transfer source term per unit length can be obtained
by multiplying the evaporation rate from a single particle,
o
d m , by the total number of particles in the control volume
[9]:
A m N S d
o
p mass
= (7)
The number of particles per unit volume, N
p
, can be
expressed as:
3
6
p
d
p
d
N
t
o
= (8)
The momentum coupling source term due to the reverse
effect of particles can be written as [20]:
2
1
( )
2 4
p
mom p D p g d g d
d
S N A C x u u u u
t
p = ÷ ÷ ÷ (9)
The energy coupling source term for the total energy
equation evolves convective heat transfer and the work due to
particle drag (first and second terms on the right hand side,
respectively) [20].
2
( )
energy p p g d mom d
S N Ah d T T S u
;
t = ÷ ÷ + (10)
C. Friction force
The friction force per unit length between the pipe wall and
the gas phase was estimated by,
2
) (
2
g g g pipe wg
u
f
d F o p t = (11)
The friction factor, f, can be calculated from the well
known Blasius formula. In addition, the friction factor
between particles and the wall of the pipe as in [22] is,
831 . 1
0503 . 1
÷
=
p p
Fr f (12)
Where, Fr
p
= u
d
/(gd
p
)
0.5
is the particle Froude number.
D.Description of the particulate solids
The model used in the present study assumes that the solids
will be modeled as true spheres with a vastly increased surface
area to allow for the roughness and protuberance, while the
pores were envisaged as cylindrical, running from the surface
to the particle centre, as depicted in [8]. The particulate solid
properties given in Table 1 need to be determined and the
following parameters that describe the solid particles can be
developed through these properties as follows:
The sphericity, Ȥ, can be defined as the ratio between the
true surface area and the spherical surface area as given in [8]
as:
6
p sa so
d A p
; = (13)
The critical water content of solids, X
cr
, can be estimated
from the following equation;


.

\

÷ =
s sa
w cr
X
p p
p
1 1
(14)
The density of the dispersed phase, which is composed
from liquid water and solid material, can be expressed as;
) 1 ( X
sa d
+ = p p (15)
The diameters of pores are assumed to be normally
distributed around a mean value of dm with standard deviation
of ı. The following equations were developed to describe this
model mathematically [8].
The diameters of pores are assumed to be normally
distributed around a mean value of dm with standard deviation
of ı. The following equations were developed to describe this
model mathematically [8].
min
0.5
2
2 2
min
min
8 1 1
64 1 1
4 1 1
1
pore sa s
sa s pore
pore sa s
d A
d A
d A
þ
p p
p p
p p
 
= ÷ ÷

\ .
 
÷

\ .
+
 
+ ÷ ÷

\ .
(16)


.

\

÷
+
=
s
sa
sa pore
m
A
d
p
p
þ p
1
) 1 (
4
2
(17)
International Journal of Aerospace and Mechanical Engineering 5:2 2011
112
m
d þ o = (18)
Excel 2003 is used to generate the size distribution using
the built in function NORMSDIST. The obtained distribution,
shown in Fig. 1, is then fed to the computer program.
TABLE I PROPERTIES OF THE PARTICULATE SOLID [8]
Property Value for alumina
Average particle diameter, d
p
0.0001 m
Diameter of smallest pores in
solid, d
min
7×10
9
m
External solid surface area, A
so
333 m
2
/kg dry solids
Internal solid surface area, A
pore
6×10
4
m
2
/kg dry
solids
Density of solid material, ȡ
s
3700 kg/m
3
Density of average particle, ȡ
sa
1600 kg/m
3
Heat capacity of dry solid, C
ps
,
as function of temperature (K)
6.9540.2803T
0.25

11.604T
0.15
kJ/kg.K
0.0E+0 1.0E8 2.0E8 3.0E8 4.0E8
0E+0
2E+5
4E+5
6E+5
N
u
m
b
e
r
o
f
p
o
r
e
s
o
f
s
i
z
e
d
p
o
r
e
Pore diameter, d (m)
pore
Fig. 1 Pore size distribution in an individual alumina particle
E. Heat transfer
The convective heat transfer coefficient, h, was calculated
from Nusselt number, Nu, which is expressed as a function of
Reynolds number, Re
p
and Prandtl number, Pr, which are
defined as:
g
pg g
g
p
k
C
Pr ;
hd
Nu
u
u
= =
Various empirical correlations can be used to calculate the
heat transfer coefficient. The following correlations have been
tested in the present study:
 Frantz correlation [8]
The correlation was used by Radford [8] to calculate the
heat transfer coefficient in pneumatic conveying dryer.
667 0 6 1
015 0
. .
p
Pr Re . Nu = (19)
 De Brandt correlation [1, 9]
The correlation was developed for a pneumatic dryer.
667 0 3 1
16 0
. .
p
Pr Re . Nu = (20)
 Debrand correlation [23]
The correlation was developed for a pneumatic dryer.
333 . 0 15 . 1
Pr Re 035 . 0
p
Nu = (21)
 Baeyens et al. correlation [1]
The correlation was developed for a large scale pneumatic
dryer.
p
Re . Nu 15 0 = (22)
 Modified RanzMarshall correlation [9]
The correlation was developed for single droplet/wet
particle and it takes into account the resistance of the liquid
vapor around the particle to the heat transfer by spalding
number, B.
( )
7 0
333 0 5 0
1
6 0 2
.
. .
p
B
Pr Re .
Nu
+
+
= (23)
( )
fg
p g pwv
H
T T C
B
÷
= (24)
 Modified Weber correlation [5]
An additional term proportional to
0.8
Re
p
was added to
RanzMarshall correlation to account for turbulent flow.
( )
333 0 8 0 5 0
06 0 5 0 2
. .
p
.
p
Pr Re . Re . Nu + + = (25)
The heat exchange per unit length between the
surroundings and the gas phase, Q
wall
, may be given as;
( )
wall g pipe pipe wall
T T h d Q ÷ = t (26)
In the regime of welldeveloped turbulence, the relation
between the coefficient of heat transfer between inner wall of
the pipe and gas flow, h
pipe
, and friction factor, f, can be
expressed accurately with the dimensionless quantity as
reported in [22] by;
2
f
C u
h
pg g g
pipe
=
p
(27)
F. Mass transfer coefficient
In analogy to the heat transfer coefficient, h, the mass
transfer coefficient, h
m
, is calculated from Sherwood number,
Sh, which is equivalent to Nusselt number, Nu. It is often
expressed as a function of the particle Reynolds number, Re
p
,
International Journal of Aerospace and Mechanical Engineering 5:2 2011
113
and Schmidt number, Sc, which equivalent to Prandtl number,
Pr, and they defined as:
v g
g
v
p m
D
Sc ;
D
d h
Sh
p
u
= = (28)
Eqs. 1925 have been used to calculate the mass transfer
coefficient in the present study.
G.Mass transfer rate
The mass transfer in the present model is based on the two
stage drying process. In the first stage, the solid surface can be
considered to be fully wetted and the resistance to the mass
transfer is located in the gas side. The evaporation rate from
individual particle can be expressed as given in [9] as:


.

\

9
÷
9
=
g
vg w
d
vo w
p m d
o
T
p M
T
p M
d h m
2
;t (29)
Where, p
vo
and p
vg
are the partial pressures of water vapor at
the particle surface and the gas phase.
The second drying stage period starts when the particulate
surface becomes no longer wetted and evaporation must occur
from within the pores. This was assumed to occur at solid
water content, X, less than the critical solid water content, X
cr
.
Radford [8] mentioned that there are five possible
mechanisms of evaporation during this period (the fallingrate
period). The predominant mechanism at any time will depend
upon the pressure, temperature and the diameter of pore from
which evaporating is occurring. The five mechanisms are
summarized blow.
Evaporation mechanism 1. It was mentioned earlier that
the pores were envisaged as cylindrical, running from the
surface to the particle centre. Evaporation from these pores
will be aided by capillary action. The same surface tension
effects that forced the water to the surface will lower the
vapor pressure of the water, slowing its evaporation rate. The
vapor pressure reduction is described by [8] as:


.

\

÷ =
d g pore
L t
vo vo
T R d
V
p p
o 4
exp
1
(30)
The surface tension force also produces a reduced pressure
in the water immediately behind the meniscus (pressure
reduction = 4ı
t
/d
pore
) [8]. As the pore diameter decreases the
pressure reduction increases and the pressure in the water
decreases. If the pressure falls blow the vapor pressure at the
prevailing temperature, the water behind the meniscus started
to vaporize and the integrity of the meniscus will destroyed. In
the pores with a small enough diameter for this to occur, a
stable meniscus cannot form and the water contained in the
pore will not be driven to the surface by the surface tension
effects. The pore diameter at which the meniscus will break
down, d
men
, can be estimated from the following equation [8]:
vo
t
men
p P
d
÷
=
o 4
(31)
The water will evaporate from the hemispherical area at
the pore end, and the evaporation rate from pores of diameter
greater than d
men
in an individual particle can be estimated
from the following equation [8];
2
1
1
2
o
m pore pore w vg
w vo
d
d g
h d N M p
M p
m
T T
;t  
= ÷ 

9 9
\ .
(32)
Where, N
pore
is the number of pores of diameter d
pore
.
Evaporation mechanism 2. In pores of diameter less than d
men
the water will evaporate at a distance blow the surface. The
vapor must find its way to the surface by either Fick or
Knudsen diffusion flow. If the pore diameter is significantly
greater than the mean free path of water molecules at the
prevailing temperature and pressure, the vapor flow will be
described by Fick diffusion. The evaporation rate from all
bores of diameter greater than the free path of water vapor and
less than d
men
can be described as given in [24] by;
2
2
,
18
4
o
pore pore vo vg
v
d
g p B lm
d N p p
D P
m
R T Z p
t ; ÷
=
A
(33)
Where, P
B,lm
is the logmean pressure of nonH
2
O gas and is
given by;
)] /( ) ln[(
,
vo vg
vg vo
lm B
p P p P
p p
p
÷ ÷
÷
= (34)
The diffusion path length, ǻZ, can be estimated from Eq. (35)
by assuming the pores are cylindrical of constant diameter
through its entire length [8];
cr
cr
p
X
X X
d
Z
÷
= A
2
(35)
Evaporation mechanism 3. In pores of diameter comparable
with the mean free path of water vapor molecules, the vapor
flow is controlled by Knudsen diffusion rather than Fick
diffusion. The rate of evaporation from all pores of diameter
comparable with the mean free path of water vapor molecules
in an individual particle will be given by [24]:
5 . 0
3
3
) (
) (
04433 . 0
d
vg vo pore pore
d
o
RT Z
p p d N
m
A
÷
=
;
(36)
The transition pore diameter from Fick diffusion to
Knudsen diffusion was estimated by setting the evaporation
rate by Fick diffusion equal to that by Knudsen diffusion as
follows [8]:
5 . 0
,
6
10 612 . 1
R p
PT
d
lm B
d
tran
÷
× = (37)
Evaporation mechanism 4. In order to use either Fick
diffusion or Knudsen diffusion, the value of the diffusion path
length, ǻZ, must be defined. At a solid water content of X
cr
,
there will be no free water on the surface but all pores are
totally. Therefore, ǻZ will be zero and Fick and Knudsen
diffusions will yield infinite evaporation rate. As this an
unrealistic situation, it was assumed that the evaporation from
International Journal of Aerospace and Mechanical Engineering 5:2 2011
114
the pores will initially be at the rate given by Eq. (32) but with
reduced evaporation area (circular area instead of
hemispherical area) and without the effect of surface tension
on the vapor pressure [8]. Equation (32) can be rewritten to
describe the current situation as follow:


.

\

9
÷
9
=
g
vg w
d
vo w
pore pore m
d
o
T
p M
T
p M
N d h
m
4
2
4
;t
(38)
The evaporation rate form pores of diameter less than d
men
will
continue to be described by Eq. (38) until the value of
evaporation rate given by Eq. (33) or Eq. (36) is less than that
estimated by Eq. (38).
Evaporation mechanism 5. The equation for calculating the
logmean pressure of nonH
2
O gas, P
B,lm
, contains an implicit
assumption that the vapor pressure at the surface from which
evaporation is occurring, p
vo
, is less than the total gas
pressure, P. If the particle temperature is increased to a value
at which the vapor pressure is greater than the total ambient
pressure, then this assumption will be invalidated. At this
condition, the evaporation rate will be controlled by the
friction force. By assuming laminar flow of water vapor
through the pores, the evaporation rate from pores of diameter
d
pore
will be given by [8];
( ) P p
Z
N d
m
vo
wv
pore pore wv
d
o
÷
A
=
u
; tp
128
4
5 (39)
The above five mechanisms describe the evaporation rate
during the second drying stage (fallingrate period). The
appropriate evaporation mechanism to any specific pore will
depends upon the prevailing condition at the time under
consideration. The selection of the drying mechanism must be
established for each pore at the prevailing conditions.
Therefore, this aspect will be discussed in a later section of
this paper.
H. Supplementary Equations
In order to solve the above set of equations several
supplementary equations, definitions and empirical
correlations are required. These will be presented
subsequently. It should be noted that both the gas and solid
phases are mixtures and hence their thermodynamic properties
are calculated using the mixture theory.
 The volume fraction equation
1 = +
d g
o o (40)
 Mass and mole fraction of water vapor in the gas stream
( )
2
2
;
1
o
wv
wv da
wv H O o o
wv da wv H O
wv da
m M m
m y
m M m M
m m
= =
+ ÷
+
(41)
Where, M
da
and
O H
M
2
are molecular weight of dry air
and water vapor, respectively.
 Density of gas stream
The mole fraction of water vapor is used together with the
ideal gas equation to calculate the density of the gas phase as
follow:
/
g g g
P R T p = (42)
Where,
( )
2 2 2
/ ; 1
g g g H o H o H o da
R M M y M y M = 9 = + ÷ (43)
 Heat capacity of the gas stream
( )
pda wv pwv wv pg
C m C m C ÷ + = 1 (44)
 Viscosity of gas stream
( )
da O H wv O H g
y y u u u
2 2
1÷ + = (45)
 Thermal conductivity of gas stream
( )
da wv wv wv g
k m k m k ÷ + = 1 (46)
 Heat capacity of the dispersed phase
ps pw pd
C
X
C
X
X
C
+
+
+
=
1
1
1
(47)
The effect of temperature on the physical properties
of water vapor and dry air is calculated from formula given in
[25].
III. NUMERICAL METHOD
The system of equations (16) with the help of auxiliary
and supplementary equations is solved numerically using the
conservative variable formulation for the gas phase [26] and
the fourth order RoungeKutta method is used for the
dispersed phase. The conservative variable formulation is a
cell by cell iterative procedure in which the gas phase
variables are specified at the cell inlet and that at the cell exit
are sought. The average values of the gas phase variables are
then used to calculate the solid phase velocity and
temperature. The source terms are then evaluated and new
flow variables at the cell exit can be calculated. The procedure
is continued until the gas velocity no longer changes with
continued iteration. Once the solution is obtained for one cell,
the exit conditions are taken as the starting condition for the
adjacent cell and the procedure is repeated. More details about
the application of conservative variable formulation for single
phase and twophase flows can be found in [26].
The mass transfer mechanisms obviously explained can
be used to determine the evaporation from free surface water
through the evaporation from the smallest pores in the porous
particles. The evaporation rate from individual particle must
be calculated for each iteration of the conservative variable
formulation procedure. During the fallingrate period, the
evaporation rate mechanisms cannot be distributed absolutely
and must be established for each pore diameter under each set
of conditions existing at the current iteration. Figure 2 depicts
this aspect as it considered in the present model.
IV. MODEL VALIDATION
The present model was firstly validated against pneumatic
transport data (without heat or mass transfer). Hariu and
Molstad [27] measured the pressure drop in a vertical glass
tube as a function of the solid mass flow rate. The same
vertical tube, the same solid particles and the same initial
International Journal of Aerospace and Mechanical Engineering 5:2 2011
115
conditions are used to simulate the pressure drop in the
present study. In the present model as well as Arastoopour and
Gidaspow model [28], the initial void fraction or the initial
solid velocity is needed. Because these values were not
measured by Hariu and Molstad [27], the inlet gas void
fraction was assumed to be 0.955 in both simulations.
Figure 3 presents the comparison between the present
predictions, the measured pressure drop by Hariu and Molstad
[27] and the numerical predictions of Arastoopour and
Gidaspow [28]. The predicted pressure drop using the
correlation of Konno and Saito, reported in [1], is also
presented in the figure. The figure indicates that the present
model predicts the linear dependence of pressure drop with
solid mass flow rate very well compared with that of [28]. The
correlation of Konno and Saito predicts the linear variation of
pressure drop with solid mass flow rate but with higher values
than the experimentally observed and presently predicted.
Fig. 2 A flowchart indicating distribution of evaporation mechanisms according to the prevailing conditions.
International Journal of Aerospace and Mechanical Engineering 5:2 2011
116
0 7 14 21 28 35
0
500
1000
1500
2000
2500
P
r
e
s
s
u
r
e
d
r
o
p
(
p
a
)
Solid mass flow rate (kg/hr)
Correlation Ref. [1]
Experimental, Ref. [27]
Numerical, Ref. [28]
Present Prediction
Fig. 3 Comparison between predicted pressure drop in a pneumatic
conveying tube and published numerical and experimental results
For more validation of the present model,
comparisons are carried out for highspeed gassolid flow
(choked flow). Mobbs et al. [29] measured the solid velocity
for three types of solid particles in a brass pipe of Ǭ in (0.022
m) diameter and 148 ft (45.11 m) long. The conditions for the
comparison are given in Table 2. Because the particles used in
the experiment have a range of size, simulations are carried
out for the largest and the smallest particles. Figure 4 shows
the comparison between the present predictions and the
measured solid velocity for the three types of solid particles. It
can be seen from the figure that for the cases involving
polystyreneair suspension, the agreement between the
computed and the measured values is reasonably good. The
model underpredicts the solid velocity for the case involving
silicaair suspension and overpredicts it for steel shotair
suspension. The irregularity of the silica particles has similar
effect to the wall friction as it produces acceleration to the
subsonic flow [20]. Since the drag coefficient and Nusslet
number correlations used in the present model are for
spherical particles, the model underestimates the solid
velocity, as shown in Fig. 4b. The steel shot particles are
heavier and the frequency of particlewall collision is high.
These are not taken into account in the present model.
Therefore, the higher values of the predicted steel shot
velocities are probably attributed to the absence of particle
wall collision in the present model. Overall, however, the
agreement between the present model and the measured data
for highspeed gas solid flow is quite acceptable.
0.0 0.2 0.4 0.6 0.8 1.0
40
80
120
160
200
x/L
S
o
l
i
d
v
e
l
o
c
i
t
y
(
m
/
s
)
Experimental Ref. [29]
Prediction, d = 151 m
Prediction, d = 422 m
p
p
u
u
a. polystyrene
0.0 0.2 0.4 0.6 0.8 1.0
40
80
120
160
200
x/L
S
o
l
i
d
v
e
l
o
c
i
t
y
(
m
/
s
)
Experimental Ref. [29]
Prediction, d = 20 m
Prediction, d = 140 m
p
p
u
u
b. silica
0.0 0.2 0.4 0.6 0.8 1.0
0
40
80
120
x/L
S
o
l
i
d
v
e
l
o
c
i
t
y
(
m
/
s
)
Experimental Ref. [29]
Prediction, d = 178 m
Prediction, d = 500 m
p
p
u
u
c. steel shot
Fig. 4 comparisons between predicted solid velocity and
experimental data of Mobbs et al. [29]
TABLE II CONDITIONS FOR HIGHSPEED GASSOLID FLOW COMPARISONS [29]
Type ȡ
d
(kg/m
3
) d
p
(ȝm)
o
g
m (kg/s) /
o o
g
s
m m
C
pd
(J/kg.K)
Polystyrene 1058 151422 0.1040 2.328 1300
Silica 1200 20140 0.1022 1.9458 703
Steal shot 7575 187500 0.1153 1.6772 500
International Journal of Aerospace and Mechanical Engineering 5:2 2011
117
The model is then validated against pneumatic conveying
dryer data. The experimental data of Radford [8] are selected
for validation purpose. The conditions of the test cases are
given in Table 3. The pipe diameter and length are 0.2 and 6
m, respectively. Because the inlet void fraction or the solid
velocity is needed in the present model, the inlet solid velocity
was assumed to be 0.2 of the inlet gas velocity.
Several Nusselt number correlations are tested in the
present study. Figure 5 presents comparisons between the
present predictions using different correlations and the
experimental data given by previous investigations. It can be
seen from the figure that the correlation of Baeyens et al. [1]
gives the best agreement with experimental data. The outlet
gas temperature and the outlet solid moisture content was
underpredicted by about 4% and 0.6 %, respectively, when
this correlation was applied. The modified RanzMarshall
correlation [9] and the modified Weber correlation [5] are in
acceptable agreement with the experimental data. They under
predict the outlet temperature and the outlet solid moisture
content by about 5% and 20%, respectively. The figure shows
also that the De Brandt correlation [1, 9] predicts the
temperature profile fairly well while it overpredicts the solid
moisture content by about 40%. On the other hand, Frantz [8]
and Debrand [23] correlations give very poor results. Since
the correlation of Baeyens et al. [1] gives the best predictions,
this correlation is used in the present study to calculate the
heat and mass transfer coefficients.
Figure 6 shows comparisons between the present
predictions using Baeyens et al. correlation and the
experimental results of Radford (1997) under different
conditions. From this figure it can be seen that the present
model predicts the gas temperature and solid water content
very well. The figure shows also that the surface moisture is
removed in the first few millimeters of the conveying duct.
This can be attributed to the high inlet gas temperature and the
high slip velocity between phases at this region.
TABLE III CONDITIONS FOR COMPARISON TEST CASES [8]
Inlet gas
temperature
[k]
Inlet solid
temperature
[k]
Feed gas mass flow
rate kg/sec
Wet solid mass
flow rate
Water content
of solid kg/kg
dry solid
Water content
in feed gas
stream (%)
Case 1 873 300 0.5606 0.4897 0.41 3.5
Case 2 1073 300 0.4681 0.4767 0.41 4.5
Case 3 1173 300 0.5328 0.4990 0.41 4.8
0.0 2.0 4.0 6.0
300
400
500
600
700
800
900
Experimental
Modified Weber
Debrand
Frantz
Modified RanzMarshal
De Brandt
Baeyens et al.
G
a
s
t
e
m
p
e
r
a
t
u
r
e
(
k
)
Distance along the dryer (m)
0.0 2.0 4.0 6.0
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
X
(
k
g
w
a
t
e
r
/
k
g
d
r
y
s
o
l
i
d
)
Distance along the dryer (m)
a gas temperatures b water content
Fig. 5 Comparison between different heat and ass transfer correlations and experimental data (case 1)
International Journal of Aerospace and Mechanical Engineering 5:2 2011
118
a gas temperatures
b water content
Fig. 6 Comparison between present predictions and experimental
results of Ref. [8].
V.RESULTS AND DISCUSSION
The effects of inlet gas temperature, T
gi
, air mass
flow rate,
o
air
m , solid mass flow rate,
S
o
m on the axial
distribution of gas temperature, T
g
, solid Temperature, T
d
, gas
velocity, u
g
, solid velocity, u
d
, solid water content, X, and gas
water content,
O H
m
2
are studied for the parameters given in
Table 4, as shown in Figs. 710. The same solid particles and
the same pipe of Radford [8] are used in the simulations. In
general, it can be seen from the figures that the gas
temperature continuously decreases along the dryer, which in
turn results in a decrease of the gas velocity. This can be
attributed to the increase in gas density along the dryer with
the decrease in gas temperature. The rapid decrease of gas
temperature in the early stage of drying suggests that the
pneumatic conveying dryer can be used for heat sensitive
materials with higher inlet gas temperature. On entering the
drying tube, the solid velocity increases rapidly then starts to
decrease due to the decrease of gas velocity. Similar
observations were reported by Kemp et al. [7]. The particles
leave the dryer at a velocity approximately equal to that of the
gas. The solid temperature firstly decreased in the few
millimeters near the dryer inlet due to the high evaporation
rate in this region and starts to increase after the free surface
water is removed. Due to the continuous decrease of gas
temperature, the solid temperature starts to decrease again.
The maximum solid temperature occurs at a distance ranged
between 0.25 and 1 m from the tube inlet, depending upon the
inlet conditions.
TABLE IVSUMMARY OF PARAMETERS USED IN CALCULATIONS
Dry air
mass
flow rate
(kg/s)
Dry solid
mass flow
rate (kg/s)
Inlet gas
temperature (K)
Inlet
velocity slip
coefficient
0.45
*
0.15 800
*
0.1
0.75 0.35
*
1000 0.2
*
0.95 0.50 1200 0.3
*
base case, X
i
= 0.41 kg water/kg dry solid, T
di
= 300 K , m
wv
= 0.045 kg water vapor/kg dry air
Figure 7 shows the effect of changing the inlet gas
temperature with respect to the base case on the velocities,
temperatures and moisture contents for both phases. It can be
seen from this figure that the gas velocity increases as the inlet
gas temperature increases. This is due to the decrease in gas
density as the gas temperature increase. The figure also
indicates that, as the gas temperature increases the solid
temperature and the gas temperature content increase and the
solid moisture content decreases. This is due to the fact that an
increase in the gas inlet temperature results in an increase in
the temperature difference driving force and the gas velocity,
which in turn results in an increase in the heat and mass
transfer coefficients.
The effect of drying air mass flow rate on the
velocities, temperatures and moisture contents for both phases
is presented in Fig. 8. It can be seen from this figure that as
the air mass flow rate increases the gas velocity also increases.
Two major phenomena take place as the superficial inlet gas
velocity increases. The first, the particle residence time
decreases with increasing superficial inlet gas velocity
because particle velocity increases with increasing gas
velocity, as shown in Fig. 8b, and therefore the drying time
reduces. The second, increasing gas velocity increases heat
and mass transfer coefficient. Also, the total heat contained in
the gas increases with increasing the gas mass flow rate at
identical inlet gas temperature, so that the amount of heat
provided to vaporize moisture in the particles increases as the
gas mass flow rate. Both phenomena are working against each
other depending on the characteristics of gas and particles as
well as the hydrodynamic characteristics in the dryer.
Namkung and Cho [18] and Bunyawanichakul et al. [30]
reported that the drying rate increases as the inlet gas velocity
increases while Pelegrina and Crapiste [12], Kaensup et al
International Journal of Aerospace and Mechanical Engineering 5:2 2011
119
[19] and Saravnan et al. [31] reported that the drying rate
decreases as the inlet gas velocity increases. In the present
study, the drying rate increases (the solid moisture content
decreases) with increasing gas mass flow rate because the
effect of increased heat and mass transfer on drying rate
outweighs that of reduced particle residence time.
0 1 2 3 4 5 6
10
20
30
40
50
60
G
a
s
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
Tgi = 800 k
Tgi = 1000 k
Tgi = 1200 k
a Gas velocity
0 1 2 3 4 5 6
10
15
20
25
30
35
S
o
l
i
d
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
b Solid velocity
0 1 2 3 4 5 6
200
400
600
800
1000
1200
g
a
s
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
c Gas temperature d Solid temperature
0 1 2 3 4 5 6
0.05
0.10
0.15
0.20
0.25
m
(
k
g
w
a
t
e
r
/
k
g
d
r
y
a
i
r
)
Distance along the dryer (m)
w
v
e Gas water content
0 1 2 3 4 5 6
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
X
(
k
g
w
a
t
e
r
/
k
g
d
r
y
s
o
l
i
d
)
Distance along the dryer (m)
d Solid water content
Fig. 7 Effect of inlet gas temperature, T
gi
, on the axial distribution of velocities, temperatures and water content along the
dryer for both phases
0 1 2 3 4 5 6
290
300
310
320
330
340
350
S
o
l
i
d
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
0.00 0.01 0.02
290
300
International Journal of Aerospace and Mechanical Engineering 5:2 2011
120
0 1 2 3 4 5 6
10
20
30
40
50
60
70
80
G
a
s
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
mair = 0.45 kg/s
mair = 0.75 kg/s
mair = 0.95 kg/s
a Gas velocity
0 1 2 3 4 5 6
10
20
30
40
50
S
o
l
i
d
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
b Solid velocity
0 1 2 3 4 5 6
300
400
500
600
700
800
g
a
s
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
c Gas temperature
0 1 2 3 4 5 6
290
300
310
320
330
340
350
360
S
o
l
i
d
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
d Solid temperature
0 1 2 3 4 5 6
0.04
0.08
0.12
0.16
m
(
k
g
w
a
t
e
r
/
k
g
d
r
y
a
i
r
)
Distance along the dryer (m)
w
v
e Gas water content
0 1 2 3 4 5 6
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
X
(
k
g
w
a
t
e
r
/
k
g
d
r
y
s
o
l
i
d
)
Distance along the dryer (m)
f Gas water content
Fig. 8 Effect of air mass flow rate,
o
air
m , on the axial distribution of velocities, temperatures and water content along the dryer for
both phases
Figure 9 presents the effect of solid mass flow rate on
the velocities, temperatures and moisture contents for both
phases. As the solid mass flow rate increases the solids holdup
increases, resulting in an increased quantity of moisture in the
system. This enhances the saturation of the drying gas, as
shown in Fig. 9e, and decreases the gas temperature, as shown
in Fig. 9c. As a result, the driving force for heat and mass
transfer decreased, which eventually decreases the drying rate.
Qualitative similar observations are reported by Saravanan et
al. [31]. As the gas temperature decreases the gas density
increases which in turn resulting in a decrease of both solid
and gas velocities, as shown in Fig. 9a and b.
The present model as well as any other twofluid
model requires the definitions of the inlet parameters. The
solids velocity or the solids void fraction is essential to
calculate the gas phase void fraction and velocity by means of
International Journal of Aerospace and Mechanical Engineering 5:2 2011
121
0 1 2 3 4 5 6
15
20
25
30
35
40
G
a
s
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
ms= 0.15 kg/s
ms = 0.30 kg/s
ms= 0.50 kg/s
a Gas velocity
0 1 2 3 4 5 6
5
10
15
20
25
30
S
o
l
i
d
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
b Solid velocity
0 1 2 3 4 5 6
300
400
500
600
700
800
g
a
s
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
c Gas temperature
0 1 2 3 4 5 6
290
300
310
320
330
S
o
l
i
d
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
d Solid temperature
0 1 2 3 4 5 6
0.04
0.08
0.12
0.16
0.20
m
(
k
g
w
a
t
e
r
/
k
g
d
r
y
a
i
r
)
Distance along the dryer (m)
w
v
e Gas water content
0 1 2 3 4 5 6
0.15
0.20
0.25
0.30
0.35
0.40
X
(
k
g
w
a
t
e
r
/
k
g
d
r
y
s
o
l
i
d
)
Distance along the dryer (m)
f Solid water content
Fig. 9 Effect of solid mass flow rate,
o
s
m , on the axial distribution of velocities, temperatures and water content along the dryer for
both phases
void fraction equation, Eq. (40), and the gas phase continuity
equation, respectively. The inlet solids velocity and the inlet
solid void fraction are related by means of the solid phase
continuity equation. However, these parameters are not
usually measured. The effect of inlet slip coefficient, S
vo
=
u
di
/u
gi
, on the velocities, temperatures and moisture contents
for both phases is presented in Fig. 10. It can be seen from this
figure that the inlet solid velocity contributed in predicting the
other properties of pneumatic conveying dryer. It also proves
the necessity of measuring the void fraction or solid velocity
in any experimental analysis.
International Journal of Aerospace and Mechanical Engineering 5:2 2011
122
0 1 2 3 4 5 6
15
20
25
30
35
40
G
a
s
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
Svo= 0.1
Svo= 0.2
Svo= 0.3
a Gas velocity
0 1 2 3 4 5 6
0
5
10
15
20
25
S
o
l
i
d
v
e
l
o
c
i
t
y
(
m
/
s
)
Distance along the dryer (m)
b Solid velocity
0 1 2 3 4 5 6
300
400
500
600
700
800
g
a
s
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
c Gas temperature
0 1 2 3 4 5 6
280
290
300
310
320
330
S
o
l
i
d
t
e
m
p
e
r
a
t
u
r
e
(
K
)
Distance along the dryer (m)
d Solid temperature
0 1 2 3 4 5 6
0.04
0.08
0.12
0.16
m
(
k
g
w
a
t
e
r
/
k
g
d
r
y
a
i
r
)
Distance along the dryer (m)
w
v
e Gas water content
0 1 2 3 4 5 6
0.20
0.25
0.30
0.35
0.40
X
(
k
g
w
a
t
e
r
/
k
g
d
r
y
s
o
l
i
d
)
Distance along the dryer (m)
f Solid water content
Fig. 10 Effect of inlet velocity slip coefficient, S
vo
, on the axial distribution of velocities, temperatures and water content along the
dryer for both phases.
VI. CONCLUSIONS
Onedimensional steadystate nonequilibrium two
phase model has been developed to demonstrate the drying of
porous materials in a vertical upward gassolid system. The
model takes into account the momentum, heat and mass
transfer between the continuous phase and the dispersed
phase. The model was solved numerically using the
conservative variables formulation for the gas phase and the
fourth order RoungeKutta for the dispersed phase. The model
was validated against pneumatic transport and pneumatic
drying experimental results and a good agreement was
obtained. The effects of different operating conditions on the
properties of pneumatic drying were studied. It was found that
the drying rate was increased as the inlet gas temperature or
International Journal of Aerospace and Mechanical Engineering 5:2 2011
123
the gas mass flow rate increased, while it was decreased as the
solid mass flow rate was increased. Moreover, the model can
be used to calculate the length of the drying tube for specific
outlet solids moisture content.
NOMENCLATURE
A Pipe crosssectional area [m
2
]
C
p
Specific heat [J/kg.K]
C
D
Drag coefficient []
d
pipe
Pipe diameter [m]
d
p
Particle diameter [m]
d
pore
Pore diameter [m]
D
v
diffusivity of water in air [m
2
/s]
g Gravity acceleration [m/s
2
]
h Heat transfer coefficient [W/m
2
.K]
H Enthalpy [J/kg]
H
fg
Latent heat of evaporation [J/kg]
h
m
mass transfer Coefficient [m/s]
k Thermal conductivity [W/m.s]
m Mass fraction []
M Molecular weight [kg/kmol]
m
p
Mass of single wet particle [kg]
o
d
m Evaporation rate from single particle [kg/s]
o
di
m Evaporation rate by mechanism i [kg/s]
P Total gas pressure [N/m
2
]
P
vg
Partial pressure of water vapor in gas stream
[N/m
2
]
p
vo
water vapor pressure at T
d
[N/m
2
]
R Gas constant [J/kg.K]
T Temperature [K]
S
v0
Ratio between solid velocity and gas velocity
(u
d
/u
g
)
t
d
Residence time [s]
u Velocity [m/s]
V
L
Molar volume of water [m
3
/kmol]
x Axial distance along the dryer [m]
X Solid phase moisture content [kg/kg dry solid]
y Mole fraction []
GREEK LETTERS
Į Void fraction []
ȡ Density [kg/m
3
]
ȝ Viscosity [kg/m.s]
9 Universal gas constant [J/kmol.K]
ı Standard deviation [m]
ı
t
Surface tension [N/m]
Ȥ Surface area shape factor (Sphericity)
Subscripts
cr Critical
d Dispersed phase
da Dry air
g Gas phase
H
2
O Water vapor
i Inlet
s Solid material
w Water
wv Water vapor
REFERENCES
[1] Baeyens J., Gauwbergen D. V. and Vinckier I., 1995, "Pneumatic
Drying: the Use of LargeScale Data in a Design Procedure", Powder
Tech., Vol. 83, pp. 139148.
[2] Borde I. and Levy A., 2006, "Pneumatic and Flash Drying", in
Handbook of Industrial Drying, 3rd Ed., (Mujumdar A. S. Ed.), CRC
Press, New York.
[3] Moyers C. G. and Baldwin G. W., 1997, "Psychrometry, Evaporative
Cooling, and Solids Drying", in Perry’s Chemical Engineers’ Handbook,
7th Ed. (Eds. Robert H. Perry R. H., Green D. W. and Maloney J. O.),
McGrawHill, Inc.
[4] Thorpe G.R., Wint A. and Coggan G.C., 1973, "The Mathematical
Modelling of Industrial Pneumatic Driers", Transactions of the
Institution of Chemical Engineers, Vol. 51, pp. 339–348.
[5] Kemp I. C., Bahu R. E. and Pasley H. S., 1994, "Model Development
and Experimental Studies of Vertical Pneumatic Conveying Dryers",
Drying Tech., Vol. 12, pp. 13231340.
[6] Kemp I. C and Oakley D. E., 1997, "Simulation and Scaleup of
Pneumatic Conveying and Cascading Rotary Dryers", Drying Tech.,
Vol. 15, pp. 16991710.
[7] Kemp I. C., Oakley D. E. and Bahu R. E., 1991, "Computational Fluid
Dynamics Modelling of Vertical Pneumatic Conveying Dryers", Powder
Tech., Vol. 65, pp. 477484.
[8] Radford R. D., 1997, "A Model of Particulate Drying in Pneumatic
Conveying Systems", Powder Tech., Vol. 93, pp. 109126.
[9] Levy A. and Borde I., 1999, "Steady State One Dimensional Flow Model
for a Pneumatic Dryer", Chem. Eng. Processing, Vol. 38, pp. 121130.
[10] Skuratovsky I., Levy A. and Borde I., 2003, "TwoFluid Two
Dimensional Model for Pneumatic Drying", Drying Tech., Vol. 21, pp.
16451668.
[11] Skuratovsky I., Levy A. and Borde I., 2005, "TwoDimensional
Numerical Simulations of the Pneumatic Drying in Vertical pipes",
Chem. Eng. Processing, Vol. 44, pp. 187192.
[12] Pelegrina A. H. and Crapiste G. H., 2001, "Modelling the Pneumatic
Drying of Food Particles", J. Food Eng., Vol. 48, pp. 301310.
[13] Narimatsu C. P., Ferreira M. C. and Feire J. T., 2007, "Drying of Coarse
Particles in a Vertical Pneumatic Conveyor", Drying Tech., Vol. 25, pp.
291302.
[14] Fyhr C. and Rasmuson A., 1997, "Mathematical Model of Pneumatic
Conveying Dryer", AICHE Journal, Vol. 43, pp. 28892902.
[15] Fyhr C. and Rasmuson A., 1997, "Steam Drying of Wood Ships in
Pneumatic Conveying Dryers", Drying Tech., Vol. 15, pp. 17751785.
[16] Alvarez P. I., Vega R. and Blasco R., 2005, "Cocurrent Downflow
Fluidized Bed Dryer: Experimental Equipment and Modeling", Drying
Tech., Vol. 23, pp. 14351449.
[17] NamKung W. and Cho M., 2004, "Pneumatic Drying of Iron Ore
particles in a Vertical tube", Drying Tech., Vol. 22, pp. 877891.
[18] Kaensup W., Kulwong S. and Wongwises S., 2006, "A SmallScale
Pneumatic Conveying Dryer of Rough Rice", Drying Tech., Vol. 24, pp.
105113.
[19] Kaensup W., Kulwong S. and Wongwises S., 2006, "Comparison of
Drying Kinetics of Paddy Using a Pneumatic Dryer with and without a
Cyclone", Drying Tech., Vol. 24, pp. 10391045.
[20] Hamed M. H., 2005," Choked GasSolid TwoPhase Flow in Pipes", J.
Eng. Applied Science, Faculty of Engineering, Cairo University, Vol.
52, pp. 961980.
[21] Kladas, D. D. and Deorgiou, D. P., 1993, "A Relative Examination of
CD–Re Relationships used in Particle Trajectory Calculations", Trans.
of ASME, J. of Fluids Engineering, Vol. 115, pp. 162165.
[22] Han, T., A. Levy, and Y. Peng, 2000, "Model for Dilute GasParticle
Flow in ConstantArea Lance with Heating and Friction ", Powder
Technology, Vol. 112, pp. 283288.
[23] Debrand S., 1974, "Heat Transfer during a Flash Drying Process", Ind.
Eng. Chem., Process Des. Develop., Vol. 13, pp. 396404.
International Journal of Aerospace and Mechanical Engineering 5:2 2011
124
[24] Welty J. R., Wicks G. E. and Wilson R. E., 1984, "Fundamentals of
Momentum, Heat, and Mass Transfer", 3rd Ed., John Wiley & Sons,
New York.
[25] Reynolds W. C., 1979, "Thermodynamic Properties in SI: Graphs,
Tables and Computational Equations for Forty Substances", Published
by the Department of Mechanical Engineering, Stanford University,
Stanford CA 94305.
[26] Crowe C., Sommerfeld M. and Tsuji Y., 1998, "Multiphase Flow with
Droplets and Particles", CRC Press, Florida, USA.
[27] Hariu O. H. and Molstad M. C., 194, "Pressure Drop in Vertical Tubes
in Transport of Solids by Gases", Industrial and Engineering Chemistry,
Vol. 41, pp. 11481160.
[28] Arastoopour H. and Gidaspow D., 1979, "Vertical Pneumatic Conveying
Using Four Hydrodynamic Models", Ind. Eng. Chem. Fundam., Vol. 18,
pp. 123130.
[29] Mobbs F. R., Bowers H. M., D. M. Riches D. M., and Cole B. N., 1969
70, "Influence of Particle Size Distribution on the highSpeed flow of
GasSolid Suspensions in a Pipe", Proc Instn Mech Engrs, Vol. 184, Pt
3C, paper 9.
[30] Bunyawanichakul P., Walker G. J., Sargison J. E. and Doe P. E., 2007,
"Modelling and Simulation of Paddy Grain (Rice) Drying in a simple
Pneumatic Dryer", Biosystems Eng., Vol. 96, pp. 335344.
[31] Saravanan B., Balsubramaniam N. and Srinivasakannan C., 2007,
"Drying Kinetics in a Vertical GasSolid System", Chem. Eng. Tech.,
Vol. 30, pp. 176183.
International Journal of Aerospace and Mechanical Engineering 5:2 2011
125
it is important to notice that the difference between the present model and that of Radford [8] is that Radford's model assumes constant pressure along the duct and neglects the velocity difference between the two phases. Smom and Senergy are mass. which were performed in a vertical upward pneumatic dryer. Narimatsu et al. the model was not verified with experimental results. The gas phase is mixture of water vapor and dry air. respectively. therefore.The total energy equation for the gas phase can be written as: d dx g g gu g A H g 2 ug 2 Qwall (3) u d2 2 S energy g ug Ag S mass H wv Where. The model which was developed by Hamed [20] has been modified and used here based on the following assumptions: The flow is onedimensional and steady. They found that. This implies that any mass. heat or momentum transfer between particles is insignificant to the transfer between individual particles and the gas stream. momentum and energy coupling source terms. according to their assumption. The mass transfer process was obtained by the comprehensive single particle model described by Radford [8]. [16] have studied numerically and experimentally the drying process in a vertical downward pneumatic dryer. A. it was noticed that the morphology of particles (porous or non porous) did not influence in the air temperature profiles. Also. measurements of the temperature and pressure profiles as well as the final moisture content were carried out. Smass. Furthermore. The particles are spherical in shape. The model formulations are similar to that of Levy and Borde [9] but the present model allows for higher temperature.The equation of motion for a particle in a gas is given by: du d2 dx 3 g CD 2 dd p g d u g ud u g ud u u fp d d d (4) 2g 1 111 . . The present paper concerns with a onedimensional model for a pneumatic drying of porous particles. The irregular movement and the non spherical shape of the wood ships were accounted by measuring drag and heat transfer coefficients. in the low range of air flow rates. The model assumes that the solid will be conveyed as discrete particles and that heat and mass transfer occur from individual particles. Unlike the above studies. An opposite effect was observed at higher velocities. The gas pressure in the pneumatic drying process is around atmospheric pressure. a quasione dimensional situation has been considered. To validate the model.The mass balance equation for the gas phase may be written as: d ( g g u g A ) S mass (1) dx . the pressure drop under drying conditions is higher than that under transport conditions. Fyhr and Rasmuson [1415] presented a more complex model for a pneumatic dryer considering a distribution of particle sizes for steam drying of wood chips. and the measurements of heat transfer coefficient indicted that the maximum value of heat transfer coefficient occurred at the velocity of minimum pressure drop. However. They assumed that the internal resistance does not control the mass and energy transfer between solid particles and air.MATHEMATICAL MODELING To formulate the suggested model. and the predictions agreed well with the experimental results. The model includes a comprehensive twodimensional model for single particle drying of single wood ship and onedimensional plug flow was assumed. [13] investigated numerically and experimentally the drying process of porous alumina and solid glass particles in a vertical dryer.by [1719]. The model is concerned with the twophase flow of gas and particles through a vertical pipe. II. The model of Levy and Borde [9] assumes that the drying process stops and particle breakup occurs if the saturation pressure inside the wet core is greater than the ambient pressure. The model was for non shrinkage spherical particle and steady state onedimensional flow. particle breakup may be occurred if the temperature inside the wet core exceeds 100oC at any point within the duct. All particles interaction is ignored in the model. The model was for onedimensional incompressible flow and the internal resistance did not control the heat and mass transfer. Dry solids were used in heat transfer experiments.The momentum equation for the gas phase can be expressed as: d ( dx g g 2 ug A ) A Fwg dP dx S mom g g gA (2) S mass u d . Governing Equations Based on the above mentioned assumptions the governing equations for the gas and dispersed phases are derived according to the basic laws of fluid mechanics as follows: .International Journal of Aerospace and Mechanical Engineering 5:2 2011 and heat transfer coefficients. Alvarez et al. Some experimental works on the pneumatic dryer were given also.
831 (12) Where. running from the surface to the particle centre.International Journal of Aerospace and Mechanical Engineering 5:2 2011 where. The equation for particle temperature. Fwg d pipe CD 24 Re p 24 Re 0. assuming the temperature is uniform throughout the particle can be written as: dT d 2 u d m pC pd d p h (T g T d ) dx (5) m d H fg o The residence time of the particles at the gas phase was calculated as given in [9] as: dt d dx 1 ud (6) Aso sa dp 6 (13) B.5 f 2 g ( g ug )2 (11) Re p 1 Re p 400 Re p 1 400 3 10 5 The friction factor. The following equations were developed to describe this model mathematically [8]. 8 d min A pore 64 2 2 d min A pore 4 d min A pore 1 sa S mass N p md A o (7) The number of particles per unit volume. Np. fp 1. by the total number of particles in the control volume [9]: o The critical water content of solids. X cr 1 w sa 1 s (14) The density of the dispersed phase. Re p g d p ug ud / g is the particle Reynolds number. while the pores were envisaged as cylindrical. can be defined as the ratio between the true surface area and the spherical surface area as given in [8] as: where. Xcr. respectively) [20]. 2 (10) S energy N p Ah d p (T g T d ) S mom u d C.Description of the particulate solids The model used in the present study assumes that the solids will be modeled as true spheres with a vastly increased surface area to allow for the roughness and protuberance.Friction force The friction force per unit length between the pipe wall and 1 sa 1 s 2 0. Frp = ud/(gdp)0. In addition.646 p 0. which is composed from liquid water and solid material. as depicted in [8].0503Frp 1. can be calculated from the wellknown Blasius formula. CD is the drag coefficient and it is calculated as given in [21] as: the gas phase was estimated by. D. Coupling between phases An important concept in the analysis of twophase gassolid flows is to consider the mutual effect between the two phases. m d .5 1 s (16) 1 s 1 sa 1 dm A pore 4 sa (1 2 ) 1 sa s (17) 112 . The mass transfer source term per unit length can be obtained by multiplying the evaporation rate from a single particle. The particulate solid properties given in Table 1 need to be determined and the following parameters that describe the solid particles can be developed through these properties as follows: The sphericity. can be expressed as: Np 6 d d 3 p (8) The momentum coupling source term due to the reverse effect of particles can be written as [20]: 2 dp 1 S mom N pA CD ud ) u g ud (9) p x (u g 2 4 The energy coupling source term for the total energy equation evolves convective heat transfer and the work due to particle drag (first and second terms on the right hand side. f. The diameters of pores are assumed to be normally distributed around a mean value of dm with standard deviation of . can be expressed as.5 is the particle Froude number. the friction factor between particles and the wall of the pipe as in [22] is. . X) (15) d sa (1 The diameters of pores are assumed to be normally distributed around a mean value of dm with standard deviation of . can be estimated from the following equation. The following equations were developed to describe this model mathematically [8].
Heat transfer The convective heat transfer coefficient. can be expressed accurately with the dimensionless quantity as reported in [22] by.0E+0 Nu 1.15 Pr 0.8 was added to p RanzMarshall correlation to account for turbulent flow. d pore (m) 4.15 kJ/kg. h. Density of average particle. The obtained distribution. Fig.K 0. Pr. as function of temperature (K) 6E+5 Number of pores of size d pore Nu 4E+5 2 0. 1 Pore size distribution in an individual alumina particle Qwall d pipe h pipe Tg Twall E. may be given as. Qwall. the relation between the coefficient of heat transfer between inner wall of the pipe and gas flow.0E8 Pore diameter. Rep. shown in Fig.8 Pr 0. Nu 0.5 Re 0.0E8 2 0. Heat capacity of dry solid. dp Diameter of smallest pores in solid. Nu 0.06 Re 0. Cps.2803T 11.0E8 3. The following correlations have been tested in the present study: Frantz correlation [8] The correlation was used by Radford [8] to calculate the 113 . Aso Internal solid surface area. hpipe.333 p p (25) the (26) The heat exchange per unit length between surroundings and the gas phase.6 Re 0. Nu. Mass transfer coefficient In analogy to the heat transfer coefficient.333 Baeyens et al.667 p Debrand correlation [23] (20) Property Average particle diameter.035 Re1p.International Journal of Aerospace and Mechanical Engineering 5:2 2011 dm (18) heat transfer coefficient in pneumatic conveying dryer. Sh. which is equivalent to Nusselt number.25 The correlation was developed for a large scale pneumatic dryer. was calculated from Nusselt number.3 Pr 0. TABLE I PROPERTIES OF THE PARTICULATE SOLID [8] 0. Various empirical correlations can be used to calculate the heat transfer coefficient. Nu. which are defined as: In the regime of welldeveloped turbulence. s sa Value for alumina 0. f. Rep and Prandtl number. h pipe g Nu hd p g .15 Re p (22) Modified RanzMarshall correlation [9] The correlation was developed for single droplet/wet particle and it takes into account the resistance of the liquid vapor around the particle to the heat transfer by spalding number. Nu Excel 2003 is used to generate the size distribution using the built in function NORMSDIST. is calculated from Sherwood number.7 (23) B 2E+5 C pwv Tg H fg Tp (24) Modified Weber correlation [5] An additional term proportional to Re0.5 Pr 0. It is often expressed as a function of the particle Reynolds number. B. which is expressed as a function of Reynolds number.9540. correlation [1] (21) 3700 kg/m3 1600 kg/m3 6. is then fed to the computer program. h. 1.0001 m 7×109 m 333 m2/kg dry solids 6×104 solids m2/kg dry The correlation was developed for a pneumatic dryer. hm.16 Re1.5 0.6 Pr 0. Nu 0. and friction factor.333 p 1 B 0. 0E+0 0. dmin External solid surface area. 9] (19) The correlation was developed for a pneumatic dryer. Apore Density of solid material.667 p De Brandt correlation [1. the mass transfer coefficient.604T0.015 Re1.0E8 2. Pr g C pg u g C pg f 2 (27) kg F.
the value of the diffusion path length.lm R 0. As this an unrealistic situation. Evaporation from these pores will be aided by capillary action. Evaporation mechanism 1. there will be no free water on the surface but all pores are totally.International Journal of Aerospace and Mechanical Engineering 5:2 2011 and Schmidt number. Sc g g Dv (28) Eqs. If the pore diameter is significantly greater than the mean free path of water molecules at the prevailing temperature and pressure. temperature and the diameter of pore from which evaporating is occurring.lm p vo ln[( P p vg p vo )] p vg ) /( P (34) The diffusion path length. PB. 2 o hm d pore N pore M w pvo 1 M w pvg md1 (32) 2 Td Tg Where. It was mentioned earlier that the pores were envisaged as cylindrical. Z d p X cr X 2 X cr (35) p vo1 p vo exp 4 tVL d pore R g Td (30) Evaporation mechanism 3.Mass transfer rate The mass transfer in the present model is based on the twostage drying process. slowing its evaporation rate. running from the surface to the particle centre. In order to use either Fick diffusion or Knudsen diffusion. Evaporation mechanism 2. 1925 have been used to calculate the mass transfer coefficient in the present study. The same surface tension effects that forced the water to the surface will lower the vapor pressure of the water. The vapor pressure reduction is described by [8] as: p B . can be estimated from Eq. a stable meniscus cannot form and the water contained in the pore will not be driven to the surface by the surface tension effects. If the pressure falls blow the vapor pressure at the prevailing temperature. it was assumed that the evaporation from 114 . Therefore. G. Xcr. In pores of diameter less than dmen the water will evaporate at a distance blow the surface.5 (37) Evaporation mechanism 4. Z. and the evaporation rate from pores of diameter greater than dmen in an individual particle can be estimated from the following equation [8]. 2 o 18 d pore N pore Dv P pvo pvg md2 (33) R gT p Z p B . less than the critical solid water content. In the first stage. Sc. the vapor flow will be described by Fick diffusion. pvo and pvg are the partial pressures of water vapor at the particle surface and the gas phase. md hm 2 dp M w p vo Td M w p vg Tg (29) Where. and they defined as: d men 4 P t p vo (31) Sh hm d p Dv . The second drying stage period starts when the particulate surface becomes no longer wetted and evaporation must occur from within the pores. must be defined. Radford [8] mentioned that there are five possible mechanisms of evaporation during this period (the fallingrate period). can be estimated from the following equation [8]: Z ( RTd ) 0. the vapor flow is controlled by Knudsen diffusion rather than Fick diffusion.5 (36) The transition pore diameter from Fick diffusion to Knudsen diffusion was estimated by setting the evaporation rate by Fick diffusion equal to that by Knudsen diffusion as follows [8]: d tran 1. The predominant mechanism at any time will depend upon the pressure. dmen. As the pore diameter decreases the pressure reduction increases and the pressure in the water decreases. the solid surface can be considered to be fully wetted and the resistance to the mass transfer is located in the gas side. Pr. X.lm is the logmean pressure of nonH2O gas and is given by. (35) by assuming the pores are cylindrical of constant diameter through its entire length [8]. At a solid water content of Xcr. which equivalent to Prandtl number. Npore is the number of pores of diameter dpore. The evaporation rate from individual particle can be expressed as given in [9] as: o The water will evaporate from the hemispherical area at the pore end. The rate of evaporation from all pores of diameter comparable with the mean free path of water vapor molecules in an individual particle will be given by [24]: o md3 0. Z will be zero and Fick and Knudsen diffusions will yield infinite evaporation rate.612 10 6 PTd p B .lm 4 Where.04433 3 N pore d pore ( p vo p vg ) The surface tension force also produces a reduced pressure in the water immediately behind the meniscus (pressure reduction = 4 t/dpore) [8]. The pore diameter at which the meniscus will break down. In the pores with a small enough diameter for this to occur. Z. The evaporation rate from all bores of diameter greater than the free path of water vapor and less than dmen can be described as given in [24] by. The vapor must find its way to the surface by either Fick or Knudsen diffusion flow. In pores of diameter comparable with the mean free path of water vapor molecules. The five mechanisms are summarized blow. This was assumed to occur at solid water content. the water behind the meniscus started to vaporize and the integrity of the meniscus will destroyed.
Mass and mole fraction of water vapor in the gas stream mwv m wv o o . the evaporation rate mechanisms cannot be distributed absolutely and must be established for each pore diameter under each set of conditions existing at the current iteration. The evaporation rate from individual particle must be calculated for each iteration of the conservative variable formulation procedure.The volume fraction equation g d 1 o (40) .Heat capacity of the gas stream (38) C pg m wv C pwv 1 m wv C pda 1 y H 2O (44) The evaporation rate form pores of diameter less than dmen will continue to be described by Eq. PB. respectively. y H 2O m wv m da mwv M da mwv M da 1 mwv M H 2O (41) Where. During the fallingrate period. NUMERICAL METHOD The system of equations (16) with the help of auxiliary and supplementary equations is solved numerically using the conservative variable formulation for the gas phase [26] and the fourth order RoungeKutta method is used for the dispersed phase. The selection of the drying mechanism must be established for each pore at the prevailing conditions. P. III. These will be presented subsequently. (38) until the value of evaporation rate given by Eq. (38). (32) but with reduced evaporation area (circular area instead of hemispherical area) and without the effect of surface tension on the vapor pressure [8]. (36) is less than that estimated by Eq. By assuming laminar flow of water vapor through the pores. the same solid particles and the same initial md5 wv d 4 pore N pore Z 128 p vo P (39) wv The above five mechanisms describe the evaporation rate during the second drying stage (fallingrate period). the exit conditions are taken as the starting condition for the adjacent cell and the procedure is repeated. The appropriate evaporation mechanism to any specific pore will depends upon the prevailing condition at the time under consideration. (33) or Eq. At this condition. The source terms are then evaluated and new flow variables at the cell exit can be calculated. is less than the total gas pressure. The procedure is continued until the gas velocity no longer changes with continued iteration. Equation (32) can be rewritten to describe the current situation as follow: o g P / R gT g (42) Where. pvo. M g y H 2o M H 2o 1 y H 2o M da (43) md 4 hm 2 d pore N pore M w p vo 4 Td M w p vg Tg . The mass transfer mechanisms obviously explained can be used to determine the evaporation from free surface water through the evaporation from the smallest pores in the porous particles.Density of gas stream The mole fraction of water vapor is used together with the ideal gas equation to calculate the density of the gas phase as follow: 115 . Therefore. definitions and empirical correlations are required. . The same vertical tube. .Thermal conductivity of gas stream kg 1 mwv k da 1 1 X (46) . the evaporation rate from pores of diameter dpore will be given by [8]. IV. Once the solution is obtained for one cell. Supplementary Equations In order to solve the above set of equations several supplementary equations. contains an implicit assumption that the vapor pressure at the surface from which evaporation is occurring. Evaporation mechanism 5. H. o . The conservative variable formulation is a cell by cell iterative procedure in which the gas phase variables are specified at the cell inlet and that at the cell exit are sought. the evaporation rate will be controlled by the friction force. then this assumption will be invalidated. The average values of the gas phase variables are then used to calculate the solid phase velocity and temperature. If the particle temperature is increased to a value at which the vapor pressure is greater than the total ambient pressure. More details about the application of conservative variable formulation for single phase and twophase flows can be found in [26]. Mda and M H 2 O are molecular weight of dry air and water vapor.International Journal of Aerospace and Mechanical Engineering 5:2 2011 the pores will initially be at the rate given by Eq. MODEL VALIDATION The present model was firstly validated against pneumatic transport data (without heat or mass transfer). It should be noted that both the gas and solid phases are mixtures and hence their thermodynamic properties are calculated using the mixture theory. Hariu and Molstad [27] measured the pressure drop in a vertical glass tube as a function of the solid mass flow rate. this aspect will be discussed in a later section of this paper.Heat capacity of the dispersed phase C pd C pw C ps (47) The effect of temperature on the physical properties of water vapor and dry air is calculated from formula given in [25].lm.Viscosity of gas stream g y H 2O mwv k wv X 1 X wv da (45) . The equation for calculating the logmean pressure of nonH2O gas. Figure 2 depicts this aspect as it considered in the present model. Rg /M g.
the measured pressure drop by Hariu and Molstad [27] and the numerical predictions of Arastoopour and Gidaspow [28].International Journal of Aerospace and Mechanical Engineering 5:2 2011 conditions are used to simulate the pressure drop in the present study. Figure 3 presents the comparison between the present predictions. the inlet gas void fraction was assumed to be 0. The predicted pressure drop using the correlation of Konno and Saito. The figure indicates that the present model predicts the linear dependence of pressure drop with solid mass flow rate very well compared with that of [28]. is also presented in the figure. In the present model as well as Arastoopour and Gidaspow model [28]. reported in [1]. 116 . The correlation of Konno and Saito predicts the linear variation of pressure drop with solid mass flow rate but with higher values than the experimentally observed and presently predicted. 2 A flowchart indicating distribution of evaporation mechanisms according to the prevailing conditions. Fig. Because these values were not measured by Hariu and Molstad [27]. the initial void fraction or the initial solid velocity is needed.955 in both simulations.
steel shot Fig.0 0.2 a. [28] Present Prediction Correlation Ref.1022 0.0 c. d p = 178 m Prediction.9458 1.0 0 0. It can be seen from the figure that for the cases involving polystyreneair suspension.8 1. [29] measured the solid velocity for three types of solid particles in a brass pipe of in (0.4 0.2 0.4 0. the higher values of the predicted steel shot velocities are probably attributed to the absence of particlewall collision in the present model. 4b. the agreement between the present model and the measured data for highspeed gas solid flow is quite acceptable. [1] 200 Experimental Ref.6 0. however.6772 117 . [29] 160 120 80 40 0. The model underpredicts the solid velocity for the case involving silicaair suspension and overpredicts it for steel shotair suspension. comparisons are carried out for highspeed gassolid flow (choked flow).022 m) diameter and 148 ft (45.0 120 Experimental Ref.0 0 0 7 14 21 28 35 Solid velocity (m/s) Solid mass flow rate (kg/hr) Fig.8 1.8 1. The steel shot particles are heavier and the frequency of particlewall collision is high.6 x/L polystyrene 0.1040 0. Ref. d p = 151 m 120 80 40 Prediction. simulations are carried out for the largest and the smallest particles. d p = 500 m 40 Prediction. 3 Comparison between predicted pressure drop in a pneumatic conveying tube and published numerical and experimental results Experimental Ref.11 m) long. Because the particles used in the experiment have a range of size. [29] TABLE II CONDITIONS FOR HIGHSPEED GASSOLID FLOW COMPARISONS [29] (kg/m3) 1058 1200 7575 dp ( m) 151422 20140 187500 m g (kg/s) 0. Figure 4 shows the comparison between the present predictions and the measured solid velocity for the three types of solid particles. d p = 20 m Prediction. Since the drag coefficient and Nusslet number correlations used in the present model are for spherical particles. Mobbs et al. d p = 140 m For more validation of the present model. Overall. Ref. Type Polystyrene Silica Steal shot d 0. The irregularity of the silica particles has similar effect to the wall friction as it produces acceleration to the subsonic flow [20]. [27] Numerical. silica 0. [29] Solid velocity (m/s) 80 Prediction. the agreement between the computed and the measured values is reasonably good.328 1. the model underestimates the solid velocity. 4 comparisons between predicted solid velocity and experimental data of Mobbs et al.4 x/L 0.2 x/L b. as shown in Fig. d p = 422 m 2000 Pressure drop (pa) 1500 1000 500 0. Therefore.0 200 0.1153 o ms / m g o o Cpd (J/kg.K) 1300 703 500 2. 0. [29] Solid velocity (m/s) 160 Prediction. These are not taken into account in the present model.International Journal of Aerospace and Mechanical Engineering 5:2 2011 2500 Experimental.6 0. The conditions for the comparison are given in Table 2.
15 0. [1] gives the best predictions. This can be attributed to the high inlet gas temperature and the high slip velocity between phases at this region.20 0.5 4. The modified RanzMarshall correlation [9] and the modified Weber correlation [5] are in acceptable agreement with the experimental data.5606 0. From this figure it can be seen that the present model predicts the gas temperature and solid water content very well. Because the inlet void fraction or the solid velocity is needed in the present model.0 4.0 4.30 0.5 4. 5 Comparison between different heat and ass transfer correlations and experimental data (case 1) 118 .0 2.41 0.5328 Wet solid mass flow rate 0. 9] predicts the temperature profile fairly well while it overpredicts the solid moisture content by about 40%. Frantz [8] and Debrand [23] correlations give very poor results.10 2. Modified RanzMarshal De Brandt X (kg water/kg dry solid) 800 Gas temperature (k) 700 600 500 400 300 0.6 %. The conditions of the test cases are given in Table 3. The experimental data of Radford [8] are selected for validation purpose. when this correlation was applied.International Journal of Aerospace and Mechanical Engineering 5:2 2011 The model is then validated against pneumatic conveying dryer data. The figure shows also that the surface moisture is removed in the first few millimeters of the conveying duct. On the other hand.41 Water content in feed gas stream (%) 3.gas temperatures b. respectively.45 Modified Weber Debrand Frantz Baeyens et al. Since the correlation of Baeyens et al.4897 0. this correlation is used in the present study to calculate the heat and mass transfer coefficients.0 a.water content Fig. [1] gives the best agreement with experimental data.4681 0.41 0. TABLE III CONDITIONS FOR COMPARISON TEST CASES [8] Feed gas mass flow rate kg/sec 0.25 0. Figure 6 shows comparisons between the present predictions using Baeyens et al. Several Nusselt number correlations are tested in the present study.0 0.0 Distance along the dryer (m) 6. the inlet solid velocity was assumed to be 0.0 0. The pipe diameter and length are 0.4990 Case 1 Case 2 Case 3 Water content of solid kg/kg dry solid 0.8 900 Experimental 0. The outlet gas temperature and the outlet solid moisture content was underpredicted by about 4% and 0. They underpredict the outlet temperature and the outlet solid moisture Inlet gas temperature [k] 873 1073 1173 Inlet solid temperature [k] 300 300 300 content by about 5% and 20%.35 0.4767 0. Figure 5 presents comparisons between the present predictions using different correlations and the experimental data given by previous investigations.2 and 6 m.2 of the inlet gas velocity.0 Distance along the dryer (m) 6. respectively. correlation and the experimental results of Radford (1997) under different conditions. The figure shows also that the De Brandt correlation [1. It can be seen from the figure that the correlation of Baeyens et al. respectively.40 0.
and therefore the drying time reduces. and gas water content. Xi = 0. The figure also indicates that. solid velocity.water content Fig. gas velocity. X. The maximum solid temperature occurs at a distance ranged between 0.41 kg water/kg dry solid. Similar observations were reported by Kemp et al. The solid temperature firstly decreased in the few millimeters near the dryer inlet due to the high evaporation rate in this region and starts to increase after the free surface water is removed.50 1200 0. The same solid particles and the same pipe of Radford [8] are used in the simulations. V. the solid velocity increases rapidly then starts to decrease due to the decrease of gas velocity. This is due to the decrease in gas density as the gas temperature increase. TABLE IV SUMMARY OF PARAMETERS USED IN CALCULATIONS a. In general. Two major phenomena take place as the superficial inlet gas velocity increases. Both phenomena are working against each other depending on the characteristics of gas and particles as well as the hydrodynamic characteristics in the dryer. m H 2O are studied for the parameters given in Table 4. solid water content. [30] reported that the drying rate increases as the inlet gas velocity increases while Pelegrina and Crapiste [12]. The first.045 kg water vapor/kg dry air Figure 7 shows the effect of changing the inlet gas temperature with respect to the base case on the velocities.75 0. the total heat contained in the gas increases with increasing the gas mass flow rate at identical inlet gas temperature. solid Temperature. 8b.International Journal of Aerospace and Mechanical Engineering 5:2 2011 materials with higher inlet gas temperature. as the gas temperature increases the solid temperature and the gas temperature content increase and the solid moisture content decreases. The particles leave the dryer at a velocity approximately equal to that of the gas.1 * 0.2* 0. ud. so that the amount of heat provided to vaporize moisture in the particles increases as the gas mass flow rate. which in turn results in a decrease of the gas velocity.3 * base case. the particle residence time decreases with increasing superficial inlet gas velocity because particle velocity increases with increasing gas velocity. The second.gas temperatures Dry air Inlet Dry solid Inlet gas mass velocity slip mass flow temperature (K) flow rate coefficient rate (kg/s) (kg/s) 0. 6 Comparison between present predictions and experimental results of Ref. mwv = 0. the solid temperature starts to decrease again.45* 0. Tg. temperatures and moisture contents for both phases. [8].25 and 1 m from the tube inlet. m S on the axial distribution of gas temperature. which in turn results in an increase in the heat and mass transfer coefficients. It can be seen from this figure that the gas velocity increases as the inlet gas temperature increases. ug. [7]. air mass flow rate. as shown in Figs. Tgi. increasing gas velocity increases heat and mass transfer coefficient.35 1000 0. The rapid decrease of gas temperature in the early stage of drying suggests that the pneumatic conveying dryer can be used for heat sensitive o o 119 . Also. depending upon the inlet conditions. Due to the continuous decrease of gas temperature. Tdi = 300 K . On entering the drying tube. This is due to the fact that an increase in the gas inlet temperature results in an increase in the temperature difference driving force and the gas velocity. Td. 710.RESULTS AND DISCUSSION The effects of inlet gas temperature. mair . Kaensup et al b. It can be seen from this figure that as the air mass flow rate increases the gas velocity also increases.15 800* 0. solid mass flow rate. it can be seen from the figures that the gas temperature continuously decreases along the dryer. as shown in Fig. 8. The effect of drying air mass flow rate on the velocities.95 0. Namkung and Cho [18] and Bunyawanichakul et al. temperatures and moisture contents for both phases is presented in Fig. This can be attributed to the increase in gas density along the dryer with the decrease in gas temperature.
7 Effect of inlet gas temperature.02 0 1 2 3 4 5 Distance along the dryer (m) 6 c.Gas velocity 1200 Solid temperature (K) 350 340 330 320 310 300 290 Solid velocity (m/s) b.35 0.25 0. temperatures and water content along the dryer for both phases 120 .10 1 2 3 4 5 Distance along the dryer (m) 6 e. [31] reported that the drying rate decreases as the inlet gas velocity increases.30 0.05 0 d.01 0.10 0.Gas temperature 0.05 0 1 2 3 4 5 Distance along the dryer (m) 6 0.Solid temperature 0. 35 30 25 20 15 10 0 1 2 3 4 5 Distance along the dryer (m) 6 a.20 0.00 gas temperature (K) 1000 800 600 400 200 0 1 2 3 4 5 Distance along the dryer (m) 6 0. on the axial distribution of velocities.Gas water content d.25 mwv (kg water/kg dry air) X (kg water/kg dry solid) 0.15 0. the drying rate increases (the solid moisture content 60 50 Gas velocity (m/s) 40 30 20 10 0 1 2 3 4 5 Distance along the dryer (m) 6 Tgi = 800 k Tgi = 1000 k Tgi = 1200 k decreases) with increasing gas mass flow rate because the effect of increased heat and mass transfer on drying rate outweighs that of reduced particle residence time.20 0.Solid velocity 300 290 0. In the present study.International Journal of Aerospace and Mechanical Engineering 5:2 2011 [19] and Saravnan et al.Solid water content Fig.40 0. Tgi.15 0.
9a and b. temperatures and moisture contents for both phases. 8 Effect of air mass flow rate.25 0.35 0.Gas water content Fig. [31]. 9c.Gas temperature 0. The present model as well as any other twofluid model requires the definitions of the inlet parameters. 9e.95 kg/s 50 40 30 20 10 0 1 2 3 4 5 Distance along the dryer (m) 6 a.Gas velocity 800 b. as shown in Fig.45 kg/s Solid velocity (m/s) mair = 0. This enhances the saturation of the drying gas.08 0.Gas water content mair o . As the gas temperature decreases the gas density increases which in turn resulting in a decrease of both solid and gas velocities.30 0. As a result. as shown in Fig. the driving force for heat and mass transfer decreased.12 0.15 0. The solids velocity or the solids void fraction is essential to calculate the gas phase void fraction and velocity by means of 121 .International Journal of Aerospace and Mechanical Engineering 5:2 2011 80 70 Gas velocity (m/s) 60 50 40 30 20 10 0 1 2 3 4 5 Distance along the dryer (m) 6 mair = 0. as shown in Fig.40 0.20 0. on the axial distribution of velocities. temperatures and water content along the dryer for both phases Figure 9 presents the effect of solid mass flow rate on the velocities. f.75 kg/s mair = 0. resulting in an increased quantity of moisture in the system.Solid velocity 360 350 Solid temperature (K) gas temperature (K) 700 600 500 400 340 330 320 310 300 300 0 1 2 3 4 5 Distance along the dryer (m) 6 290 0 1 2 3 4 5 Distance along the dryer (m) 6 c.05 0 1 2 3 4 5 Distance along the dryer (m) 6 e. Qualitative similar observations are reported by Saravanan et al. which eventually decreases the drying rate.Solid temperature 0. and decreases the gas temperature.10 0.16 mwv (kg water/kg dry air) X (kg water/kg dry solid) d.04 0 1 2 3 4 5 Distance along the dryer (m) 6 0. As the solid mass flow rate increases the solids holdup increases.
The inlet solids velocity and the inlet solid void fraction are related by means of the solid phase continuity equation.Gas water content Fig.15 d.15 kg/s ms = 0. and the gas phase continuity equation. 10.International Journal of Aerospace and Mechanical Engineering 5:2 2011 40 ms= 0. Eq. 9 Effect of solid mass flow rate.Gas temperature 0.Solid water content m s . 122 . o f.Solid velocity 330 Solid temperature (K) gas temperature (K) 700 600 500 400 300 0 1 2 3 4 5 Distance along the dryer (m) 6 320 310 300 290 0 1 2 3 4 5 Distance along the dryer (m) 6 c. respectively. Svo = udi/ugi.30 kg/s ms= 0. However. It can be seen from this figure that the inlet solid velocity contributed in predicting the other properties of pneumatic conveying dryer.12 0.25 0. (40).40 0.50 kg/s Gas velocity (m/s) 35 30 25 20 15 0 1 2 3 4 5 Distance along the dryer (m) 6 30 25 20 15 10 5 0 1 2 3 4 5 Distance along the dryer (m) 6 a.30 0. temperatures and water content along the dryer for both phases void fraction equation. It also proves the necessity of measuring the void fraction or solid velocity in any experimental analysis. The effect of inlet slip coefficient.08 0. on the velocities. these parameters are not usually measured.04 0 1 2 3 4 5 Distance along the dryer (m) 6 0 1 2 3 4 5 Distance along the dryer (m) 6 e.16 0. on the axial distribution of velocities.Gas velocity 800 Solid velocity (m/s) b.20 0.35 0. temperatures and moisture contents for both phases is presented in Fig.Solid temperature 0.20 mwv (kg water/kg dry air) X (kg water/kg dry solid) 0.
heat and mass transfer between the continuous phase and the dispersed phase. It was found that the drying rate was increased as the inlet gas temperature or 123 .Gas velocity 800 Solid temperature (K) 700 600 500 400 300 0 1 2 3 4 5 Distance along the dryer (m) 6 330 320 310 300 290 280 0 1 b.Gas water content f.Solid velocity gas temperature (K) 2 3 4 5 Distance along the dryer (m) 6 c. 10 Effect of inlet velocity slip coefficient.16 mwv (kg water/kg dry air) X (kg water/kg dry solid) 0.1 Svo= 0. Svo.12 0.Solid temperature 0.25 0.Solid water content Fig.International Journal of Aerospace and Mechanical Engineering 5:2 2011 40 35 Gas velocity (m/s) 25 Svo= 0.30 0. The effects of different operating conditions on the properties of pneumatic drying were studied.2 Solid velocity (m/s) 20 15 10 5 0 30 25 20 15 0 1 Svo= 0.20 0 1 2 3 4 5 Distance along the dryer (m) 6 e.3 2 3 4 5 Distance along the dryer (m) 6 0 1 2 3 4 5 Distance along the dryer (m) 6 a.Gas temperature 0. CONCLUSIONS Onedimensional steadystate nonequilibrium twophase model has been developed to demonstrate the drying of porous materials in a vertical upward gassolid system. The model takes into account the momentum.40 d. VI. The model was validated against pneumatic transport and pneumatic drying experimental results and a good agreement was obtained. on the axial distribution of velocities.08 0. The model was solved numerically using the conservative variables formulation for the gas phase and the fourth order RoungeKutta for the dispersed phase. temperatures and water content along the dryer for both phases.35 0.04 0 1 2 3 4 5 Distance along the dryer (m) 6 0.
in Handbook of Industrial Drying.. Han.K] Standard deviation [m] Surface tension [N/m] t Surface area shape factor (Sphericity) Subscripts cr Critical d Dispersed phase da Dry air g Gas phase [18] [19] [20] [21] [22] [23] Baeyens J. C. 2001.. 2000. Drying Tech. "A Relative Examination of CD–Re Relationships used in Particle Trajectory Calculations". Chem.s] Mass fraction [] Molecular weight [kg/kmol] Mass of single wet particle [kg] Evaporation rate from single particle [kg/s] Evaporation rate by mechanism i [kg/s] Total gas pressure [N/m2] Partial pressure of water vapor in gas stream water vapor pressure at Td [N/m2] Gas constant [J/kg. NamKung W.. Vol. Eng. pp.C. Kemp I. and Y. "Pneumatic Drying: the Use of LargeScale Data in a Design Procedure". 2005. Radford R. 16991710. "Model Development and Experimental Studies of Vertical Pneumatic Conveying Dryers". E. 2006. Ind. 65. Vol. Fyhr C. and Rasmuson A... Kaensup W. I. "Steam Drying of Wood Ships in Pneumatic Conveying Dryers". Robert H. Levy A. Vol. "Simulation and Scaleup of Pneumatic Conveying and Cascading Rotary Dryers". in Perry’s Chemical Engineers’ Handbook. 52. Vol.. and Maloney J. S. Vol. and Coggan G. H. 2006. Oakley D. and Wongwises S. 16451668. and Bahu R. Process Des. 1999. Peng. Vol. 1995. 43.. pp. and Solids Drying". O. Borde I. C and Oakley D. 13231340. Levy. "Drying of Coarse Particles in a Vertical Pneumatic Conveyor". and Baldwin G. and Rasmuson A. Kemp I.. Vol. 14351449. Drying Tech.. E. 301310. "A SmallScale Pneumatic Conveying Dryer of Rough Rice". Chem. Skuratovsky I. pp. Ed. C. A.. Inc.International Journal of Aerospace and Mechanical Engineering 5:2 2011 the gas mass flow rate increased. 17751785. P. Powder Technology. Alvarez P.). Processing. 961980.. "Cocurrent Downflow Fluidized Bed Dryer: Experimental Equipment and Modeling". Kulwong S. "Model for Dilute GasParticle Flow in ConstantArea Lance with Heating and Friction ". 1974.. 44. 112. Ferreira M. and Crapiste G...s] Universal gas constant [J/kmol. 12.. "Heat Transfer during a Flash Drying Process". pp. Drying Tech. Drying Tech. 291302. "Computational Fluid Dynamics Modelling of Vertical Pneumatic Conveying Dryers".. AICHE Journal. of ASME.. 162165. Kulwong S. J. "TwoDimensional Numerical Simulations of the Pneumatic Drying in Vertical pipes".. Kaensup W. pp.. Fyhr C. pp. Develop. 1997.. "TwoFluid TwoDimensional Model for Pneumatic Drying". J. pp. NOMENCLATURE H 2O i s w wv Water vapor Inlet Solid material Water Water vapor REFERENCES A Cp CD dpipe dp dpore Dv g h H Hfg hm k m M mp o Pipe crosssectional area [m2] Specific heat [J/kg. 396404.. 2004. Applied Science. "A Model of Particulate Drying in Pneumatic Conveying Systems". Eng. "Comparison of Drying Kinetics of Paddy Using a Pneumatic Dryer with and without a Cyclone". T. 51. 22. and Pasley H. and Deorgiou. Vol. 2007. Drying Tech. pp.. Powder Tech. Vol. Transactions of the Institution of Chemical Engineers. 339–348.. 38. pp. Chem.. Levy A.. 477484.. pp. pp. "Mathematical Model of Pneumatic Conveying Dryer".. 13. pp. "The Mathematical Modelling of Industrial Pneumatic Driers". Drying Tech. D. 10391045. J. 139148. 7th Ed. and Vinckier I. 15. Narimatsu C. 187192. 1994. C. and Borde I. the model can be used to calculate the length of the drying tube for specific outlet solids moisture content. 283288.. 877891.. 28892902. pp. Debrand S. Vol. 2006. "Psychrometry. Trans... 25. pp. McGrawHill. E. "Steady State One Dimensional Flow Model for a Pneumatic Dryer". E. Skuratovsky I. while it was decreased as the solid mass flow rate was increased. CRC Press.. 1997. Green D.. and Feire J. "Modelling the Pneumatic Drying of Food Particles". 2005. 23.. S. Bahu R. H. pp.. "Pneumatic Drying of Iron Ore particles in a Vertical tube"." Choked GasSolid TwoPhase Flow in Pipes". Vol.. Vol.K] Drag coefficient [] Pipe diameter [m] Particle diameter [m] Pore diameter [m] diffusivity of water in air [m2/s] Gravity acceleration [m/s2] Heat transfer coefficient [W/m2.. Vega R.. Kladas. Vol. 21. W. Powder Tech. Vol. and Blasco R. 124 . Faculty of Engineering. Vol. pp. 93. 24. Processing. D. pp.R.. D. 1997.K] Enthalpy [J/kg] Latent heat of evaporation [J/kg] mass transfer Coefficient [m/s] Thermal conductivity [W/m.. Moyers C. Thorpe G. Eng. Vol. and Borde I.. W. (Mujumdar A. Levy A. G. Kemp I. Powder Tech.). and Levy A. H.. and Borde I. 48... 1991. Hamed M. Eng. D. and Cho M. P. Vol. Vol. Vol. "Pneumatic and Flash Drying". and Wongwises S. 105113. Drying Tech. Moreover. (Eds. Evaporative Cooling. 121130... Wint A. Drying Tech. Food Eng. pp.. New York. V. pp. 1997. 115. Gauwbergen D. 24. Vol. pp.... H. of Fluids Engineering. 3rd Ed. Vol. 109126. Cairo University. Drying Tech. 1973. 15. Perry R. 1997. 1993.K] Temperature [K] Ratio between solid velocity and gas velocity Residence time [s] Velocity [m/s] Molar volume of water [m3/kmol] Axial distance along the dryer [m] Solid phase moisture content [kg/kg dry solid] Mole fraction [] [1] [2] [3] [4] [5] [6] [7] md o [8] [9] [10] mdi P Pvg [N/m2] pvo R T Sv0 (ud/ug) td u VL x X y [11] [12] [13] [14] [15] [16] [17] GREEK LETTERS Void fraction [] Density [kg/m3] Viscosity [kg/m. 2003. 83. Pelegrina A. 2005. pp. T.
C. E. and Cole B.. Heat. 3rd Ed. 18. "Fundamentals of Momentum. Florida. 125 . 96. "Multiphase Flow with Droplets and Particles". [26] Crowe C.. and Tsuji Y. Vol.. Tech. 335344. New York. Stanford University.... R. Published by the Department of Mechanical Engineering. CRC Press. Tables and Computational Equations for Forty Substances". Vol. Industrial and Engineering Chemistry. E.. Chem. 1984. 184. and Wilson R. Eng. pp. D. pp. Fundam. [28] Arastoopour H.. Stanford CA 94305.. "Pressure Drop in Vertical Tubes in Transport of Solids by Gases". and Srinivasakannan C. Walker G. 194. Sommerfeld M. M. [25] Reynolds W.. "Vertical Pneumatic Conveying Using Four Hydrodynamic Models". and Gidaspow D. 11481160. 123130.. Sargison J. 1998. [27] Hariu O. Eng. "Thermodynamic Properties in SI: Graphs.. 176183. Vol. Pt 3C. pp. "Drying Kinetics in a Vertical GasSolid System". USA. J. 196970.. Ind.. Balsubramaniam N. E. Biosystems Eng.International Journal of Aerospace and Mechanical Engineering 5:2 2011 [24] Welty J. Wicks G. 1979. 41. "Influence of Particle Size Distribution on the highSpeed flow of GasSolid Suspensions in a Pipe".. and Mass Transfer". and Doe P. E. 2007. Vol. "Modelling and Simulation of Paddy Grain (Rice) Drying in a simple Pneumatic Dryer". M. and Molstad M. [29] Mobbs F. Chem. [31] Saravanan B. M. John Wiley & Sons.. [30] Bunyawanichakul P.. 30. Vol. C.. R. 2007. N. Bowers H.. 1979. Proc Instn Mech Engrs. H. pp. paper 9. Riches D..