International Journal of Heat and Mass Transfer 53 (2010) 4638–4650

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Three dimensional numerical modeling of simultaneous heat and moisture

transfer in a moist object subjected to convective drying
V.P. Chandra Mohan, Prabal Talukdar *
Department of Mechanical Engineering, Indian Institute of Technology Delhi, New Delhi 110016, India

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

Article history: The drying behavior of a moist object subjected to convective drying is analyzed numerically by solving
Received 9 December 2009 heat and moisture transfer equations. A 3-D numerical model is developed for the prediction of transient
Received in revised form 7 June 2010 temperature and moisture distribution in a rectangular shaped moist object during the convective drying
Accepted 17 June 2010
process. The heat transfer coefficients at the surfaces of the moist object are calculated with an in-house
computational fluid dynamics (CFD) code. The mass transfer coefficients are then obtained from the anal-
ogy between the thermal and concentration boundary layer. Both these transfer coefficients are used for
Keywords:
the convective boundary conditions while solving the simultaneous heat and mass transfer governing
Convective drying
CFD simulation
equations for the moist object. The finite volume method (FVM) with fully implicit scheme is used for
Finite volume method discretization of the transient heat and moisture transfer governing equations. The coupling between
Simultaneous heat and moisture transfer the CFD and simultaneous heat and moisture transfer model is assumed to be one way. The effect of
Optimum drying time velocity and temperature of the drying air on the moist object are analyzed. The optimized drying time
is predicted for different air inlet velocity, temperature and moisture content. The drying rate can be
increased by increasing the air flow velocity. Approximately, 40% of drying time is saved while increasing
the air temperature from 313 to 353 K. The importance of the inclusion of variable surface transfer coef-
ficients with the heat and mass transfer model is justified.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction experimental data set is not readily available for the purpose of
validation of models. Either the experimental conditions are not
Drying has been defined as the removal of volatile substances by properly described or they are not simple enough to replicate with
heated air from a moist object. During drying of moist objects simul- a numerical simulation. So, there has always been a dearth for the
taneous heat and moisture transfer occurs both inside the solid and validation of any numerical or analytical model. On the other hand,
in the boundary of the drying agent. Varieties of drying methods are the existing numerical models found in the literature are strong in
used to dry the moist objects. They are convective or direct drying the sense that they are capable of handling the simultaneous heat
[1–17], indirect or contact drying [18], di-electric or microwave dry- and moisture transfer in a moist object with many of the internal
ing [19], freeze drying [20] and natural air drying [21]. The convec- phenomena like shrinkage [7,9–11], osmosis [14] etc. But most of
tive drying is the mostly used in industries like agricultural and food these modeling works are with simple 1-D and 2-D assumption
industry, bio-oil industry, building materials, chemical/ceramic with constant surface transfer (both heat and mass) coefficients.
industry, paper industry, textile industry, nuclear waste disposal There is very limited work which uses coupled CFD and diffusion
etc. In general, in the convective drying process a considerable influ- model through a moist object with either one way [1–3] or two
ence is exerted by air flow velocity, temperature of the supplied air, way coupling.
transfer coefficient etc. Therefore, it is necessary to analyze the Convective drying problems that deal with the prediction of
influences of the above properties on drying of moist object. temperature and moisture distribution in a moist object have been
There have been a few experimental studies [6–11] of convec- reported by many researchers. Kaya et al. [1–3] developed a numer-
tive drying. The experimental data from these studies were com- ical model of heat and mass transfer during convective drying of
pared with numerical [1–4] and analytical data [13] of respective 2-D rectangular [1] and cylindrical moist objects [3] with the con-
numerical and analytical models. One shortcoming which can be vective boundary conditions and variable heat transfer coefficients
seen from the literature is that a well presented and well planned along the product surfaces. They suggested the optimum aspect
ratio, inlet and outlet locations of a rectangular cavity for optimum
heat transfer and drying rate. Hussain and Dincer [4] numerically
* Corresponding author. Tel.: +91 11 2659 6337; fax: +91 11 2658 2053.
E-mail address: prabal@mech.iitd.ac.in (P. Talukdar). predicted the temperature and moisture content distribution of a

0017-9310/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2010.06.029
 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650 4639

Nomenclature

B breadth of the moist object [m] Greek symbols

C height of the moist object [m] a thermal diffusivity [m2/s]
dt time step [s] q density of air [kg/m3]
D moisture diffusivity [m2/s] l viscosity [kg/(m s)]
D0 pre-exponential factor [m2/s] U non-dimensional moisture content
h heat transfer coefficient [W/(m2 K)] h non-dimensional temperature
hm mass transfer coefficient [m/s]
k thermal conductivity [W/(m K)] Subscripts
L length of the moist object [m] m mass, mean
Le Lewis number n normal to surface
M moisture content [kg/kg of dry air] s coordinate along the surface
n index 0 initial condition
p pressure [N/m2] 1 supplied air condition
T temperature [K]
t time [s] Superscripts
u, v, w velocities in X, Y and Z direction [m/s] * non-dimensional quantities
x, y, z distance along X, Y and Z axis
X, Y, Z coordinate axis

2-D cylindrical moist object. Dutta et al. [5] investigated experi- because the boundary layer developed and subsequent separation
mentally the drying behavior of a spherical object and validated of flow behind an object is different depending on whether it is a 2-
their 1-D numerical model with the experimental data. A mathe- D or 3-D object. This motivates the present work with a 3-D moist
matical model of 1-D simultaneous heat and mass transfer was pro- object using a simultaneous heat and mass transfer model which
posed by Wang and Brennan [6–8]. The effect of shrinkage during overcomes the general assumptions mentioned above.
convective drying was analyzed by Simal et al. [9–11]. Velic et al. This study consists of the following essential parts: (i) simula-
[12] investigated experimentally the convective drying of apple in tion of the external flow and temperature fields by an in-house
laboratory conditions and investigated the influence of airflow CFD source code, (ii) determination of convective heat transfer
velocities on drying kinetics and heat transfer coefficient. An ana- coefficient as a function of space (X, Y, Z), (iii) determination of
lytical solution of mass transfer equation with shrinkage was pro- the convective mass transfer coefficients at each location (X, Y, Z)
posed by Hernandez et al. [13] and their solution was validated using the analogy between thermal and concentration boundary
with 1-D numerical solution. Baroni and Hubinger [14] investigated layers, (iv) expressing the variable transfer coefficients by curve fit-
the drying of moist object with osmosed and fresh onions. Fernando ting equations, (v) incorporation of these curve fit equations to the
et al. [15] analyzed the surface hardened or case hardened layers convective boundary conditions of a numerical model of 3-D
leading to restricted moisture movement through resulting surface simultaneous heat and mass transfer, (vi) solution of the numerical
hardened layers and retarded drying rates. Evaluation of convective model (mentioned in (v)) for determination of temperature and
heat transfer coefficient of various crops during drying was ana- moisture content distribution as a function of time and space.
lyzed experimentally by Akpinar [17]. Murugesan et al. [21] con-
ducted an experimental and numerical simulation of natural 2. Mathematical model
convective drying with building materials. Talukdar et al. [22,23]
studied experimentally and numerically combined heat and mass 2.1. Modeling of external flow and temperature field
transfer processes through building materials.
Most of the drying models discussed above have simplifying Fig. 1 presents the schematic illustrations of the problem do-
assumptions. The usual assumptions are either: (i) the constant main. The moist object to be dried is kept inside a rectangular
surface transfer and diffusion coefficient, (ii) Dirichlet boundary channel (0.6  0.2  0.2 m) and exposed to the flow of dry hot air
condition instead of a convective boundary condition and (iii) sim- through the channel. The moist object is placed centrally at the
plified 1-D or 2-D geometry. Usually, the reason behind these
assumptions is to make the model equations simpler and thus eas-
ier to solve. In a heat and mass transfer numerical model with con-
Y
vective boundary condition, the knowledge of the surface transfer
coefficients is required. Considering constant transfer coefficient as
assumed by many of the authors may lead to poor prediction of the
X
thermal performance. In convective drying, the boundary surfaces
Z
are subjected to convective heat and mass transfer with a flowing
Moist object 0.2m
stream of dry and hot air. This necessitates a mathematical model
with convective (Robbins) boundary condition instead of a Dirich- T∞ Outlet
U∞
let (known boundary temperature and moisture content) or Neu-
mann (known flux or adiabatic) boundary conditions. Inlet
As seen from the literature, no simulations are carried out to
predict the rate of heat and mass transfer within 3-D moist objects. 0.2m
Since the transfer coefficients could also be reasonably different
0.6 m
depending on whether it is a 2-D and 3-D geometry, the extrapola-
tion of 2-D to 3-D results is not always recommended. This is Fig. 1. Moist object is subjected to convective drying.
4640 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650

middle of the channel so that it gets maximum exposure to the air dent thermo-physical properties of the material, (ii) negligible
flow. The governing partial differential equations for the forced shrinkage or deformation of the material during drying, (iii) negli-
convection motion of a drying fluid in a 3-D geometry are the mass, gible heat generation inside the moist object and (iv) negligible
momentum and energy conservation equations. The solution do- radiation effects. Under the above assumptions, the following gov-
main for the CFD simulations is the channel without the moist ob- erning 3-D heat and mass transfer equations can be written:
ject. Interior boundaries are placed around the moist object to
separate it out from the CFD simulation.
1 @T @ 2 T @ 2 T @ 2 T
¼ þ þ ð8Þ
Considering the flow as incompressible, the continuity equation a @t @x2 @y2 @z2
is 1 @M @ 2 M @ 2 M @ 2 M
¼ þ 2 þ 2 ð9Þ
@u @ v @w D @t @x2 @y @z
þ þ ¼0 ð1Þ
@x @y @z The initial conditions are,
The momentum equations with constant properties are Tðx; y; z; 0Þ ¼ T 0 and Mðx; y; z; 0Þ ¼ M o
  !
@u @u @u @p @2u @2u @2u The boundary conditions are:
q u þv þw ¼ þl þ þ ð2Þ
@x @y @z @x @x2 @y2 @z2 @Tð0; y; z; tÞ @Mð0; y; z; tÞ
  ! k ¼ hðT  T 1 Þ; D ¼ hm ðM  M1 Þ
@v @v @v @p @2v @2v @2v @x @x
q u þv þw ¼ þl þ þ ð3Þ at x ¼ 0; 0 6 y 6 C and 0 6 z 6 B
@x @y @z @y @x2 @y2 @z2
  ! @TðL; y; z; tÞ @MðL; y; z; tÞ
@w @w @w @p @2w @2w @2w k ¼ hðT  T 1 Þ; D ¼ hm ðM  M 1 Þ
q u þv þw ¼ þl þ 2 þ 2 ð4Þ @x @x
@x @y @z @z @x2 @y @z at x ¼ L; 0 6 y 6 C and 0 6 z 6 B
The energy equation with constant properties is @Tðx; 0; z; tÞ @Mðx; 0; z; tÞ
k ¼ hðT  T 1 Þ; D ¼ hm ðM  M 1 Þ
! @y @y
@T @T @T @2T @2T @2T at y ¼ 0; 0 6 x 6 L and 0 6 z 6 B
u þv þw ¼a þ þ ð5Þ
@x @y @z @x2 @y2 @z2 @Tðx; C; z; tÞ @Mðx; C; z; tÞ
k ¼ hðT  T 1 Þ; D ¼ hm ðM  M1 Þ
In the above equations, a is thermal diffusivity and l is the dy- @y @y
namic viscosity. Simulations are carried out for three different inlet at y ¼ C; 0 6 x 6 L and 0 6 z 6 B
velocities (0.1, 0.2 and 0.3 m/s). The objective of performing the @Tðx; y; 0; tÞ @Mðx; y; 0; tÞ
k ¼ hðT  T 1 Þ; D ¼ hm ðM  M1 Þ
CFD simulation is to calculate the heat transfer coefficient around @z @z
the moist object placed inside the channel. The heat transfer coef- at z ¼ 0; 0 6 x 6 L and 0 6 y 6 C
ficient is independent of the imposed temperature at the inlet, wall @Tðx; y; B; tÞ @Mðx; y; B; tÞ
of the channel and the boundaries of the moist object as long as the k ¼ hðT  T 1 Þ; D ¼ hm ðM  M1 Þ
@z @z
properties are assumed to be constant. In the simulation, an inlet
at z ¼ B; 0 6 x 6 L and 0 6 y 6 C
temperature of 333 K and wall temperature of 300 K is assumed.
The boundaries of the moist object are also set at 300 K. In the above equations, a is the thermal diffusivity in m2/s and D
A CFD source code FASTEST3D based on the finite volume meth- is the moisture diffusivity in m2/s and it is taken as a function of
od (FVM) is used to solve these equations. The grid file is generated temperature. It is obtained from the Arrhenius equation [4] as
with a commercial package ANSYS ICEM CFD. From the tempera-
D ¼ D0 expð1119=TÞ ð10Þ
ture fields obtained, the local convective heat transfer coefficient
h is determined using the balance between conductive and convec- where D0 is the pre-exponential factor, 2.41  107 m2/s [28].
tive heat transfer.

@T 
k ¼ hðT s  T m Þ ð6Þ
@ns
where, n is the normal vector at the surface s and Tm is the bulk
mean temperature. After the convective heat transfer coefficient,
h, is determined, the convective mass transfer coefficient hm is cal-
culated using the analogy between the thermal and concentration
boundary layers,
 
DLen
hm ¼ h ð7Þ
k
where, D is moisture diffusivity, a is the thermal diffusivity, k is the
thermal conductivity of the drying fluid and Le (=a/D) is the Lewis
number representing a relative measure of the thermal and concen-
tration boundary layer thickness and for most applications, n is ta-
ken as 1/3 [1–3].

2.2. Modeling of the internal temperature and moisture fields of the

object

A mathematical model is developed to analyze heat and mass

transfer through diffusion inside the object being dried. The fol- Fig. 2. (a) Velocity and (b) temperature contour around the object at XY plane at
lowing assumptions are considered: (i) moisture content indepen- Z = 0.01 m, 0.3 m/s.
 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650 4641

(a)
z=0 and 0.02m
30 z=0.002 & 0.018m

Heat transfer coeff, h(W/m2K)

z=0.004 & 0.016m
25 z=0.006 & 0.014m
z=0.008 & 0.012m
20 z=0.01m
Surface facing inlet
15

10

0
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02
Y distance in m

(b) 12 z=0 & 0.02 m

z=0.002 & 0.018 m
Heat transfer coeff, h(W/m2K)

z=0.004 & 0.016 m

10
z=0.006 & 0.014 m
z=0.008 & 0.012 m
8
z=0.01 m
Surface facing
6
outlet
4

0
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02
Y distance in m

40
(c) z=0 & 0.02 m
Heat transfer coeff, h(W/m 2K)

35 z=0.002 & 0.018 m

30 z=0.004 & 0.016 m

z=0.006 & 0.014
25 z=0.01 m

20 Top/Bottom
surfaces
15
10
5
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
X distance in m

(d) 40
Heat transfer coefficient, h(W/m K)
2

35
y=0.01 m
30 y=0.008 & 0.012 m
y=0.006 & 0.014 m
25
y=0.004 & 0.016 m
South/North
20 surfaces y=0.002 & 0.018 m
y=0 & 0.02 m
15
10

5
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
X distance in m

Fig. 3. Variation of heat transfer coefficient at (a) surface facing inlet, (b) surface facing outlet, (c) top/bottom surfaces and (d) south/north surfaces.
4642 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650

3. Numerical techniques 4. Results and discussion

The transport equations for external fluid flow and heat transfer 4.1. Results of CFD simulation
are solved by FASTEST3D [24,25], CFD code. FASTEST3D (Flow
Analysis by Solving Transport Equations Simulating Turbulence) The results of the CFD simulations are presented in this part.
is a 3-D flow prediction code. The code has been discussed in the Three CFD simulations are carried out for inlet velocities of 0.1,
previous works and hence, it is not repeated here. The grid file is 0.2 and 0.3 m/s. It is seen from the simulations, that the flow is
generated with a commercial package ANSYS ICEM CFD. steady and symmetric for the inlet velocity of 0.1 and 0.2 m/s. It
A Fortran 90 code is developed based on the finite volume dis- slightly deviates from the symmetricity at the inlet velocity of
cretization for predicting the internal heat and moisture transfer 0.3 m/s. The flow is susceptible to unsteadiness as we further in-
distribution inside the object. A fully implicit scheme is used for crease the inlet velocity. As the aspect ratio (length of the moist ob-
the time variable. At the end of the discretization process a set of ject along the flow direction/width of the moist object
algebraic equations are generated for both heat and moisture perpendicular to the flow direction) of the object decreases the
transfer. These simultaneous algebraic equations are solved using critical Reynolds number at which the flow becomes unsteady also
the Gauss–Siedel method. An iterative process at each time step decreases. For a moist object of aspect ratio 1, the flow becomes
is executed before proceeding to the next time step. Curve fitting unsteady even at a lower velocity than the case considered here.
equations are generated to define the surface transfer coefficient It is seen that at an inlet velocity of 0.3 m/s, the flow is unsteady
as a function of locations. The diffusion coefficient is taken as a var- for aspect ratio of 1. The critical Reynolds number at which the
iable property with respect to temperature while solving the mass flow becomes unsteady for a cubical object is given in the work
transfer equation. The heat and mass transfer equations are simul- of Saha [26] and our results closely agree his number. This is also
taneously solved. important to mention here that our present results do not include
The coupling between the CFD code and the diffusion code is via higher inlet velocity (higher Reynolds number). Since the flow field
the surface transfer coefficient. This coupling is one way as the ef- is unsteady for a Reynolds number larger than the critical Reynolds
fect of the conditions of the moist object over the temperature and number (case corresponding to velocity of 0.3 m/s), a one way cou-
vapor density of the drying fluid is considered to be very small. pling might give a large amount of error. The flow and temperature

(a) 35
Heat transfer coefficient, W/(m 2 K)

30
velocity = 0.1 m/s
25 velocity = 0.2 m/s
velocity = 0.3 m/s
20

15

10

0
0 0.004 0.008 0.012 0.016 0.02
Distance along Z axis, m

(b) 9

8
Heat transfer coefficient, W/(m2K)

7 velocity = 0.1 m/s

velocity = 0.2 m/s
6
velocity = 0.3 m/s
5

0
0 0.004 0.008 0.012 0.016 0.02
Distance along Z axis, m

Fig. 4. Variation of heat transfer coefficient at the boundary of the moist object facing to the inlet (a) and facing to the outlet (b) along Z direction at Y = 0.01 m.
 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650 4643

field will undergo vortex shading behind the object which will Since the trends of the velocity and temperature contours are
change the heat transfer coefficient continuously with time espe- same for all the cases of inlet velocity, only one case is presented.
cially in the boundary behind the flow field. The effect of heat Fig. 2(a) shows the velocity contours around the object in XY plane
transfer coefficient assuming it to be independent of time for these and at the middle of Z plane for an inlet velocity of 0.3 m/s. The
cases should be looked into before one present results of tempera- temperature profile around the object at XY plane (Z = 0.01 m) is
ture and moisture content distribution of the object. The other pos- shown in Fig. 2(b). The velocity and temperature contours are seen
sibility is to solve a fully coupled transient CFD and diffusion to be symmetric without any separation at the inlet velocity of 0.1
problem which has the obvious limitation of large computational and 0.2 m/s (not shown here). But at an inlet velocity of 0.3 m/s,
time. Before doing such a modeling, it is also important to see both the profiles slightly deviate from its symmetricity which
whether it is worth to make that effort relative to the accuracy that can be seen from Fig. 2(a) and (b).
can be improved. Considering these factors, in the present work, It is seen from the Fig. 2(b) that the temperature is compara-
we are not going beyond the steady regime of flow. tively lower behind the object compared to the other locations

Table 1
Variations of minimum, maximum and average heat transfer coefficient.

Surfaces (see Fig. 3) Heat transfer coefficient, W/(m2 K)

0.1 m/s 0.2 m/s 0.3 m/s
Min Max Average Min Max Average Min Max Average
South 3.49 36.41 5.85 4.08 44.02 7.13 3.72 46.42 8.03
North 3.48 36.8 5.85 4.07 44.24 7.13 3.78 46.58 8.03
Facing inlet 10.54 28.31 14.07 14.44 33.25 19.24 14.32 43.10 23.07
Facing outlet 2.02 10.8 2.73 2.54 11.62 4.15 2.36 12.49 4.88
Top 3.15 36.80 5.31 3.69 42.2 6.37 3.54 43.72 7.05
Bottom 3.15 36.80 5.31 3.69 42.2 6.37 3.54 43.72 7.05

(a)
310

309

308
Temperature, K

307
CVs 15x12x12
306 CVs 20x16x16
CVs 30x24x24
305

304

303

302
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Distance along X axis, m

(b)
320

318

316 dt = 100 s
Temperature, K

dt = 50 s
314 dt = 20 s
dt = 10 s
312
dt = 5 s
310

308

306
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Distance along X axis, m

Fig. 5. (a) Grid and (b) time independency test.

4644 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650

around the object. This makes the heat and mass transfer coeffi- Fig. 3(c) shows the variation of heat transfer coefficient at top/
cient lower in the boundary facing to the outlet of the channel bottom surfaces. The variation of heat transfer coefficient is shown
resulting a higher moisture content of the moist object in the re- in X direction with various Z locations. Fig. 3(d) shows the variation
gion closer to that boundary which can be seen later in this section. of heat transfer coefficients at the south and north surfaces of the
Fig. 3(a) shows the distribution of heat transfer coefficient at moist object. Here, variations are shown in X direction at various Y
the surface facing the inlet of the duct for an inlet velocity of values.
0.1 m/s. The distribution of heat transfer coefficient at the surface It is observed that the heat transfer coefficient is sufficiently
facing the outlet of the duct is shown in Fig. 3(b). In both these sur- higher at the boundary facing the inlet (Fig. 3(a)) because of the
faces the heat transfer coefficient varies along both Y and Z direc- immediate contact with the hot air. The heat transfer coefficient
tions. The plots are drawn between heat transfer coefficient and Y reaches a maximum value of 28.31 W/m2 K at the edges of X = 0,
directions for different Z locations. It is also to be mentioned here Y = 0 and X = 0, Y = 0.02 m and gradually decreases up to 14.1 W/
that two plots are clubbed together from both sides as the values of m2 K. On the other hand, the heat transfer coefficients are quite
heat transfer coefficient are symmetric about the Z axis. This has low at the boundary facing the outlet because of less exposure with
been done in Fig. 3(c) and (d) also. the hot air. The distribution of heat transfer coefficient at the top,
bottom, south and north boundary faces are comparable to each
other because of the orientation of the object to the flow direction.
The distribution also follows an exponentially decreasing trend
Table 2
which is because of the boundary layer developed. The heat trans-
Physical properties of moist object.
fer coefficient data for inlet velocity of 0.2 and 0.3 m/s also give
Properties approximately the similar trend of variations (not shown here).
Thermal conductivity, k 0.589 W/(m K) [28] Fig. 4(a) gives an idea how heat transfer coefficient increases at
Thermal diffusivity, a 3.488  107 m2/s [28] the boundary facing the inlet of the channel with increase in inlet
Initial moisture content, M0 7.196 kg/kg of dry air [1] velocity. The heat transfer coefficient are plotted along Z direction
Initial temperature, T0 290 K
Moisture content in supplied air, M1 0.196 kg/kg of dry air [1]
and at the center of the Y axis (Y = 0.01 m) for inlet velocity of 0.1–
0.3 m/s. The air flow velocity increases the heat transfer coefficient

(a)
333

328

323
Temperature, K

318

313
Experimental data by Wang et al. [7]
308
Numerical data with variable h
303
Numerical data with constant h
298

293
0 500 1000 1500 2000 2500 3000 3500 4000
Drying time, s

(b) 1
Non dimensional moisture content, Φ

0.8
Numerical data with variable h

Experimental data by Velic et al. [12]

0.6
Experimental data by Sarsawadia et al. [27]
Numerical data with constant h
0.4

0.2

0
0 5000 10000 15000 20000 25000 30000
Drying time, s

Fig. 6. Comparison of experimental data at 333 K and 0.3 m/s with (a) predicted temperature and (b) predicted moisture content.
 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650 4645

around the moist object almost proportionally because of in- increase in air flow velocity because of the increasing magnitude of
creased convection with increasing air flow velocity. As expected, forced convection.
the heat transfer coefficient variation shows a symmetric distribu-
tion because of uniform incoming flow. Fig. 4(b) shows variations 4.2. Results from heat and mass transfer model
of the heat transfer coefficient at the surface facing outlet and at
the same location of Y and Z as in Fig. 4(a). The heat transfer coef- Simulations for heat and mass transfer are carried out inside the
ficient values for the velocities of 0.1 and 0.2 m/s show symmetric rectangular moist object with the heat transfer coefficient calcu-
characteristics, but become asymmetric at 0.3 m/s. The asymmet- lated from the CFD simulation at the velocities of 0.1, 0.2 and
ric behavior at inlet velocity of 0.3 m/s is in consistent with the 0.3 m/s as described in the previous section. Air is considered to
velocity and temperature plots shown in Fig. 2. be the drying fluid since it is the most common practice. The moist
Table 1 presents quantitatively the convective heat transfer object is subjected to convective drying and the effect of drying
coefficients for all the three velocities considered. The values vary performance is analyzed at various temperatures ranging from of
in a range 2.02–46.58 W/m2 K for different inlet velocities. The 313 to 353 K (40–80 °C). The size of the moist object is taken as
average heat transfer coefficient for each boundary is calculated L = 0.08 m (X direction), C = 0.02 m (Y direction) and B = 0.02 m (Z
by taking a weighted average of the local heat transfer coefficient. direction). Initial conditions of the moist object are considered as
The maximum average heat transfer coefficient is obtained at the T0 = 290 K and M0 = 7.196 kg/kg of dry air [1].
surface facing inlet of the moist object for all the velocities and is First a grid independent test (Fig. 5(a)) is performed for three
maximum for maximum inlet velocity (23.07 W/m2 K for 0.3 m/s) different combinations of control volumes: 15  12  12,
considered in this work. On the other hand, the surface facing 20  16  16 and 30  24  24. It is found that the percentage dif-
the outlet of the moist object shows the lowest average heat trans- ference between the grid size 15  12  12 and 20  16  16 is
fer coefficient (varies from 2.73 to 4.88 W/m2 K for different veloc- 0.0006% and between 20  16  16 and 30  24  24 is 0.0004%.
ities) for the reason discussed earlier. The average heat transfer As there is no significant difference observed in these three cases,
coefficient is increased at all the surfaces of the moist object with the grid with 15  12  12 control volumes is taken for this analy-

(a)
0.5
Non-dimensional temperature, θ

Constant h, z*=0
0.45
Variable h, z*=0
Constant h, z*=0.0375
0.4
Variable h, z*=0.0375
Constant h, z*=0.1125
0.35
Variable h, z*=0.1125

0.3

0.25

0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Non-dimensional distance along X axis, x*

(b)
0.2
Non-dimensional moisture content, Φ

0.18

0.16

0.14

0.12 Constant h, z*=0

Variable h, z*=0
0.1
Constant h, z*=0.0375
0.08 Variable h, z*=0.0375
Constant h, z*=0.1125
0.06 Variable h, z*=0.1125

0.04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Non-dimensional distance along X axis, x*

Fig. 7. Effect of heat transfer coefficient on temperature at t* = 0.027 (a) and moisture content distribution (b) at t* = 1.09 (b). The plots are taken at a velocity of 0.1 m/s and
along the center line at y* = 0.125 at different z*.
4646 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650

sis. Similarly a time independent test is carried out with different agreement is better when a variable heat transfer coefficient is used
time steps dt (10, 20, 50, 100 and 200 s) and it is shown in in the numerical simulations. Numerical models with the assump-
Fig. 5(b). The average percentage error for temperature (at a final tion of constant heat transfer coefficient may underestimate or
time of 1000 s) between the time steps 10 and 200 is 0.1915%, overestimate the drying time.
20 and 200 is 0.0915%, 50 and 200 is 0.0312% and 100 and 200 is The results are plotted hereafter in terms of non-dimensional
0.0112%. It is seen that the difference between the time step 100 numbers to present in a generalized manner.
and 200 is quite less. Consideration of a small time step of 10 s The various non-dimensional numbers are defined below.
makes the computation very expensive for predicting the drying Non-dimensional temperature h ¼ TTT 0
1 T 0
, non-dimensional time
time. Hence, a time step of 100 s is preferred which is reasonably t* = at/L2, non-dimensional distance x* = x/L, y* = y/L and z* = z/L.
accurate and economical. The thermo-physical properties used in Fig. 7(a) shows the effect of constant (average) and variable heat
the analysis are tabulated in Table 2. transfer coefficient on non-dimensional temperature distribution
Fig. 6 shows the validation of the present work with experimen- at y* = 0.125, t* = 0.027 and for an inlet velocity of 0.1 m/s at vari-
tal data of Wang et al. [7], Velic et al. [12] and Sarsavadia et al. [27]. ous z* locations. The temperatures are maximum near to the
The comparison of predicted temperature (Fig. 6(a)) and moisture boundary (x* = 0) facing the inlet of the channel as it gets the max-
content (Fig. 6(b)) with experimental data is shown for inlet tem- imum heat from the flowing air stream. The other boundary
perature and velocity of 333 K and 0.3 m/s respectively. Fig. 6(a) (x* = 1) facing the outlet is at a lower temperature because the heat
and (b) also shows the influence of constant and variable heat transfer coefficient is quite low behind the object as discussed pre-
transfer coefficient on simulated temperature and non-dimen- viously with Fig. 3(b). The temperature in the middle portion is rel-
sional moisture content respectively. The non-dimensional mois- atively lower than both the boundaries and minimum at about
ture content (U) is calculated as, x* = 0.75 with variable heat transfer coefficient. Comparing the
plots with different Z location, it is seen that the temperature dis-
M  M air
U¼ ð11Þ tributions are higher at the surface of the moist object and gradu-
M 0  Mair
ally decreases towards the interior portion of the moist object
where, M0 is the initial moisture content. The numerical simulations (z* = 0 at the surface and 0.125 at the central plane). The tempera-
are showing good agreement with the experimental data. The ture distribution is symmetrical in nature in the XZ plane and

(a) 1

0.99
Non-dimensional temperature, θ

0.98

0.97

0.96

velocity = 0.1 m/s

0.95
velocity = 0.2 m/s
0.94 velocity = 0.3 m/s

0.93
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Non-dimensional distance along X axis, x*

(b)
0.8
Non-dimensional moisture content, Φ

0.7

0.6

0.5

velocity = 0.1 m/s

0.4
velocity = 0.2 m/s
0.3 velocity = 0.3 m/s

0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Non-dimensional distance along X axis, x*

Fig. 8. (a) Temperature and (b) moisture content distribution of moist object at y* = 0.125 and z* = 0.125 at t* = 0.27.
 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650 4647

hence the values are shown only for half portion. The temperature towards x* = 1. This is attributed to the treatment of diffusion coef-
with variable heat transfer coefficient are slightly higher than the ficient D as a function of temperature (see Eq. (10)). As the temper-
temperature with constant heat transfer coefficient for the location ature is comparatively higher closer to the boundary at x* = 1, the D
x* < 0.25 and this trend is reversed as x* increases. The difference in value decreases which in turn reduces the moisture transfer from
temperature values for constant and variable heat transfer coeffi- the object to the air flow. This results in an increase of moisture
cient observed is 2 K. content although a low moisture content is expected as this part
The consideration of constant and variable heat transfer coeffi- is close to the boundary at x* = 1.
cient has shown more influence in the moisture content results. In The effect of inlet velocity on non-dimensional temperature and
Fig. 7(b), non-dimensional moisture content distribution at moisture content is discussed in Fig. 8. Fig. 8(a) shows the non-
y* = 0.125 and at t* = 1.09 for an inlet velocity of 0.1 m/s is plotted dimensional temperature distribution along a center line in X
with various Z locations. The moisture distribution is approxi- direction (y* = 0.125 and z* = 0.125) of the moist object at
mately equal for both constant and variable heat transfer coeffi- t* = 0.27 for three different inlet velocity. The temperature gradi-
cient when x* < 0.18. As x* is greater than 0.18, considerable ents are lower for higher velocity of 0.3 m/s. The maximum differ-
variation is found in the moisture content distribution for constant ence in the temperature between the velocity of 0.1 and 0.3 m/s is
and variable heat transfer coefficient. The percentage difference is around 2 K near the boundary facing the outlet of the channel. The
less than 4% at x* < 0.18. But, the percentage difference increases non-dimensional moisture content variation is shown in Fig. 8(b)
up to 24% between the results with constant and variable heat for the same location, time and inlet velocities as for Fig. 8(a).
transfer coefficient when x* > 0.18. This is a significant difference For all the inlet velocity considered, the moisture content is mini-
and fully justifies the importance of a CFD simulation for finding mum near to the boundary facing the inlet because of higher mass
the transfer coefficients which locally vary. The distribution of transfer coefficient and gradually increases towards the middle
moisture content along X axis (for a particular z*) is almost oppo- and then slightly increases near to the boundary facing the outlet.
site to the trend of temperature distribution shown in Fig. 7(a). The maximum difference between the moisture content is 1 kg/kg
This is because the air flow is having a low moisture content of dry air for the inlet velocity range of 0.1–0.3 m/s.
(dry) and at a high temperature than the object. Comparing the
trend of the moisture content with the temperature as a mirror
image, it is seen that the moisture content is not decreasing much

Fig. 9. Temperature distributions around the object at (a) t* = 0.027, (b) t* = 0.054 Fig. 10. Moisture content distribution around the object at (a) t* = 0.27, (b)
and (c) t* = 0.27 for an inlet velocity of 0.3 m/s. The isotherms are plotted in t* = 0.817 and (c) t* = 1.36 for an inlet velocity of 0.3 m/s. The moisture content
z* = 0.125 plane. distributions are plotted in z* = 0.125 plane.
4648 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650

To have a clear view of the temperature and moisture content are relatively low towards the upstream side facing the air flow
distribution inside the object, contours are presented in Figs. 9 and high in the opposite side.
and 10. Fig. 9(a–c) shows the temperature contours at z* = 0.125 Fig. 11(a) shows the non-dimensional drying time variation of
plane, for an inlet velocity of 0.3 m/s and at a non-dimensional moist object at different inlet temperature (313–353 K) of air with
time of 0.027, 0.054 and 0.27, respectively. The temperature inside a uniform inlet velocity of 0.3 m/s. The moisture content presented
the object increases with an increase in the drying time as a result in this figure is the average moisture content of the whole object.
of convection heat transfer from the drying air to the object fol- This figure shows that the inlet temperature of the air affects the
lowed by diffusion through the object. Fig. 10(a–c) shows the con- moisture content of the moist object. Also shows that there is no
tours of moisture content inside the moist object for the same period when the drying rate is constant which is in agreement with
locations and conditions as in Fig. 9 and at t* = 0.27, t* = 0.817 the results of Baroni and Hubinger [14] and Simal et al. [9–11]. In
and t* = 1.36. The moisture content inside the object decreases food products, the drying time is defined as the time required to
with an increase in (drying) time since the moisture content of condition the moist object to a moisture content of almost 5% of
the drying air is lower than the moisture content of the object. its initial moisture content [28]. Fig. 11(b) shows the variation of
As seen the temperature and moisture content distributions are non-dimensional drying time with inlet air temperature for differ-
not symmetrical compared to those observed for the constant con- ent drying levels (5%, 10% and 15% of its initial average moisture
vective heat and mass transfer coefficients [21]. content) and for the same inlet velocity as in Fig. 11(a). The drying
Higher temperature and moisture content gradients are ob- time is reduced when the drying temperature is increased. These
tained at the region closer to the inlet because these are at the up- plots give an idea about the drying time for a certain inlet air tem-
stream side of the drying air. Moisture is transferred as vapor from perature. The drying time at different temperatures of 313, 323,
the surface of the moist object to the flowing dry air. This happens 333, 343 and 353 K are 607 min (t* = 1.985), 552 min (t* = 1.805),
due to the difference in vapor pressure at the surface of the moist 505 min (t* = 1.65), 467 min (t* = 1.527) and 435 min (t* = 1.423)
object and the dry air flowing above it. The moisture content in the respectively to reach an average moisture content of 5% to its ini-
moist object reduces with continuous drying process. It can be also tial moisture content value. While increasing drying temperature
seen that the surface moisture content during drying decreases from 313 to 353 K, it is observed that the drying time is saved by
sharply with the increase in drying time. The moisture contents approximately 40%.

(a)
1
Non-dimensional moisture content, Φ

0.9
0.8
0.7 T = 313 K
0.6 T = 323 K
0.5 T = 333 K
T = 343 K
0.4
T = 353 K
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3
Non-dimensional drying time, t*

(b)
2
Non-dimensional drying time, t*

1.6

1.2

0.8
5% of moisture content
0.4 10% of moisture content
15% of moisture content
0
310 320 330 340 350 360
Temperature of air, K

Fig. 11. (a) Variation of moisture content with drying time for different temperatures and (b) variation of drying time with temperature for different amount of drying.
 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650 4649

(a) 1

Non-dimensional moisture content, Φ

0.9
0.8
0.7
0.6 Velocity = 0.1 m/s

0.5 Velocity = 0.2 m/s

Velocity = 0.3 m/s
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3
Non-dimensional drying time, t*

(b)
2.5
Non-dimensional drying time, t*

1.5

5% of moisture content
0.5 10% of moisture content
15% of moisture content
0
0.05 0.1 0.15 0.2 0.25 0.3
Velocity of air, m/s

Fig. 12. (a) Variation of moisture content with drying time for different inlet velocities and (b) variation of drying time with inlet velocity for different amount of drying.

The non-dimensional average moisture content with non- transfer are calculated using CFD simulations with an in-house
dimensional drying time at different air flow inlet velocities at an CFD source code. The following major conclusions are drawn:
inlet temperature of 333 K are shown in Fig. 12(a). The figure
shows that the air flow rate has a significant effect on drying rates  With the increase of the air flow velocity, heat transfer coeffi-
of moist object. With the increase of the air flow velocity, the time cient is increased. The value of heat transfer coefficient is ran-
required to achieve certain moisture content is decreased. This is ged between 2.0 and 46.6 W/m2 K for the velocity range 0.1–
because, the drying rate of a material depends on how quickly 0.3 m/s considered in this work. The average heat transfer coef-
the moisture can be convected into the surrounding air followed ficient is large at left or surface facing inlet and minimum at the
by the diffusion of vapor from the inner region. Fig. 12(b) shows surface facing the outlet.
the drying time variation of the moist object with air velocity to  The result with constant heat transfer coefficient misinterprets
a moisture content of 5%, 10% and 15% of its initial moisture con- the exact drying mechanism. Results with variable heat transfer
tent value. It is seen that for an air velocity of 0.1, 0.2 and 0.3 m/ coefficient show reasonably better agreement with the experi-
s, the drying time is respectively 724 min (t* = 2.367), 599 min mental results taken from literatures.
(t* = 1.96) and 505 min (t* = 1.65) to reach an average moisture  A constant rate drying period (drying rate is constant in time) is
content of 5% to its initial moisture content value. It can be seen not detected at different air drying temperatures and different
that the increasing the air velocity by three times, drying rate air flow velocities.
can be increased approximately by 43%.  It is observed that with the increase in the air temperature the
drying rate is increased. Approximately 40% of drying time is
5. Conclusion saved while increasing the air temperature from 313 to 353 K.
 Air flow velocity shows important effect on drying mechanism.
A 3-D heat and mass transfer model is developed to study the Increasing the air velocity by three times, the drying rate can be
transient behavior of temperature and moisture content inside a increased approximately by 43%.
moist object. The drying performance of a moist object is analyzed
with the effect of air flow velocity and temperature. The variable From the obtained results, it can be concluded that the pro-
transfer coefficients required in the 3-D model of heat and mass posed heat and mass transfer model with consideration of variable
4650 V.P. Chandra Mohan, P. Talukdar / International Journal of Heat and Mass Transfer 53 (2010) 4638–4650

heat transfer coefficient, variable diffusion coefficient and convec- [15] W.J.N. Fernando, A.L. Ahmad, S.R.A. Shukor, Y.H. Lok, A model for constant
temperature drying rates of case hardened slices of papaya and garlic, J. Food
tive boundary condition is capable of predicting reasonably accu-
Eng. 88 (2008) 229–238.
rate temperature and moisture content distribution of a moist [16] K. Dworeckia, A. Slezak, B. Ornal-Wasik, S. Wasik, Evolution of concentration
object at different temperature and velocity of air. field in a membrane system, J. Biochem. Biophys. Method 62 (2005) 153–162.
[17] E.K. Akpinar, Evaluation of convective heat transfer coefficient of various crops
in cyclone type dryer, Energy Convers. Manage. 46 (2005) 2439–2454.
[18] Jian-Hua Yan, Wen-Yi Deng, Xiao-Dong Li, Fei Wang, Yong Chi, Sheng-Yong Lu,
References Ke-Fa Cen, Experimental and theoretical study of agitated contact drying of
sewage sludge under partial vacuum conditions, Drying Technol. 27 (2009)
[1] A. Kaya, O. Aydın, I. Dincer, Numerical modeling of heat and mass transfer 787–796.
during forced convection drying of rectangular moist objects, Int. J. Heat Mass [19] C.M. McLoughlin, W.A.M. McMinn, T.R.A. Magee, Microwave drying of
Transfer 49 (2006) 3094–3103. pharmaceutical powders, Inst. Chem. Eng. Trans. IChemE 78 (Part C) (2000)
[2] A. Kaya, O. Aydın, I. Dincer, Heat and mass transfer modeling of recirculating 90–96.
flows during air drying of moist objects for various dryer configurations, [20] Hongwei Wu, Zhi Tao, Guohua Chen, HongwuDeng, Guoqiang Xu, Shuiting
Numer. Heat Transfer A 53 (2008) 18–34. Ding, Conjugate heat and mass transfer process within porous media with
[3] A. Kaya, O. Aydın, I. Dincher, Numerical modeling of forced convection drying dielectric cores in microwave freeze drying, Chem. Eng. Sci. 59 (2004) 2921–
of cylindrical moist objects, Numer. Heat Transfer A 51 (2007) 843–854. 2928.
[4] M.M. Hussain, I. Dincer, Two-dimensional heat and moisture transfer analysis [21] K. Murugesan, K.N. Seetharamu, P.A.A. Narayana, A one dimensional analysis
of a cylindrical moist object subjected to drying: a finite-difference approach, of convective drying of porous materials, Heat Mass Transfer 32 (1996) 81–88.
Int. J. Heat Mass Transfer 46 (2003) 4033–4039. [22] P. Talukdar, O. Stephen Olutmayin, F. Olalekan Osanyintola, Carey J. Simonson,
[5] S.K. Dutta, V.K. Nema, R.K. Bharadwaj, Drying behavior of spherical grains, Int. An experimental data set for benchmarking 1-D, transient heat and moisture
J. Heat Mass Transfer 31 (4) (1988) 855–861. transfer models of hygroscopic building materials. Part I: experimental facility
[6] N. Wang, J.G. Brennan, A mathematical model of simultaneous heat and and material property data, Int. J. Heat Mass Transfer 50 (2007) 4527–4539.
moisture transfer during drying of potato, J. Food Eng. 24 (1993) 47–60. [23] P. Talukdar, Olalekan F. Osanyintola, Stephen O. Olutmayin, Carey J. Simonson,
[7] N. Wang, J.G. Brennan, Changes in structure, density and porosity of potato An experimental data set for benchmarking 1-D, transient heat and moisture
during dehydration, J. Food Eng. 24 (1993) 61–76. transfer models of hygroscopic building materials. Part II: experimental
[8] N. Wang, J.G. Brennan, Thermal conductivity of potato as a function of facility and material property data, Int. J. Heat Mass Transfer 50 (2007)
moisture content, J. Food Eng. 17 (1992) 153–160. 4915–4926.
[9] S. Simal, C. Rossello, A. Berna, A. Mulet, Heat and mass transfer model for [24] P. Talukdar, F.V. Issendorff, D. Trimis, C.J. Simonson, Conduction–radiation
potato drying, Chem. Eng. Sci. 49 (22) (1994) 3739–3744. interaction in 3-D irregular enclosures using the finite volume method, Heat
[10] S. Simal, A. Mulet, J. Tarrazo, C. Rossello, Drying models for green peas, Food Mass Transfer 44 (2008) 695–704.
Chem. 55 (2) (1995) 121–128. [25] I. Demirzic, M. Peric, Finite volume method for prediction of fluid flow in
[11] S. Simal, C. Rossello, A. Berna, A. Mulet, Drying of shrinking cylinder-shaped arbitrarily shaped domains with moving boundaries, Int. J. Numer. Methods
bodies, J. Food Eng. 37 (1998) 423–435. Fluids 10 (1990) 771–790.
[12] D. Velic, M. Planinic, S. Tomas, M. Bili, Influence of airflow velocity on kinetics [26] A.K. Saha, Three dimensional numerical study of flow and heat transfer from a
of convection apple drying, J. Food Eng. 64 (2004) 97–102. cube placed in a uniform flow, Int. J. Heat Fluid Flow 26 (2006) 80–94.
[13] J.A. Hernandez, G. Pavon, M.A. Garcia, Analytical solution of mass transfer [27] P.N. Sarsavadia, R.L. Sawhney, D.R. Pangavhane, S.P. Singh, Drying behaviour of
equation considering shrinkage for modeling food-drying kinetics, J. Food Eng. brined onion slices, J. Food Eng. 40 (1999) 219–226.
45 (2000) 1–10. [28] A.S. Mujumdar, Handbook of Industrial Drying, third ed., Published by CRC
[14] A.F. Baroni, M.D. Hubinger, Drying of onion: effects of pretreatment on press, Taylor and Francis group, 2006.
moisture transport, Drying Technol. 16 (9&10) (1998) 2083–2094.