International Journal of Heat and Mass Transfer 55 (2012) 54365448

Contents lists available at SciVerse ScienceDirect

International Journal of Heat and Mass Transfer

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

Numerical study of mixed convection within porous square cavities using Bejans
heatlines: Effects of thermal aspect ratio and thermal boundary conditions
D. Ramakrishna a, Tanmay Basak b, S. Roy a, I. Pop c,
a
Department of Mathematics, Indian Institute of Technology Madras, Chennai 600036, India
b
Department of Chemical Engineering, Indian Institute of Technology Madras, Chennai 600036, India
c
Faculty of Mathematics, University of Cluj, R-3400 Cluj, CP 253, Romania

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

Article history: The present numerical study deals with mixed convection ows within square enclosures lled with por-
Received 3 August 2011 ous media. The inuence of various thermal boundary conditions on bottom and side walls based on ther-
Accepted 18 April 2012 mal aspect ratio (A) is investigated for a wide range of parameters (1 6 Re 6 100, 0.015 6 Pr 6 7.2,
Available online 3 July 2012
105 6 Da 6 103 and 103 6 Gr 6 105). A penalty nite element method with bi-quadratic elements has
been used to investigate the results in terms of streamlines, isotherms and heatlines and average Nusselt
Keywords: numbers. Lid driven effect is dominant at low Darcy number (Da = 105), whereas buoyancy driven effect
Penalty nite element method
is dominant at high Darcy numbers (Da = 104 and Da = 103) for Re = 1. Asymmetric pattern is observed
Mixed convection
Square cavity
in isotherms and heatlines for Re = 100. It is found that thermal gradient is high at the center of the bot-
Thermal aspect ratio tom wall for A = 0.1 due to large dense heatlines at that zone and that is low for A = 0.9 irrespective of Re,
Streamlines Pr and Gr. Overall heat transfer rates are higher for A = 0.1 compared to other thermal aspect ratios
Heatlines (A = 0.5, A = 0.9) irrespective of Darcy number, Prandtl number and Reynolds number.
 2012 Elsevier Ltd. All rights reserved.

1. Introduction A few earlier investigations have been conducted on mixed con-

The investigations on transport phenomena through saturated
porous media are substantially increased in recent years due to
its numerous practical applications encountered in various elds
of engineering and natural science. Among these investigations,
natural and forced convection studies occupied major position.
The combined effect of natural and forced convection is commonly
referred as mixed convection. Mixed convection ow within por-
ous cavities may occur in many technical applications such as geo-
thermal engineering [1], energy recovery [2], packed bed reactors
[3], electronic cooling [4], chemical reactions [5] etc. The governing
non-dimensional parameters for mixed convection in a cavity lled
with uid saturated porous medium are Darcy number (Da), Gras-
hof number (Gr), Reynolds number (Re) and Prandtl number (Pr).
Note that, Gr and Re represent the strength of the natural and
forced convection ow effects, respectively. In addition, another
dimensionless parameter, Richardson number (Ri), may be dened
as Ri = Gr/Ren characterizes the mixed convection ow. A compre-
hensive review on the fundamentals of the convective ow in por-
ous media can also be found in the literatures [614].
Corresponding author.
E-mail addresses: dramakrishnahcu@gmail.com (D. Ramakrishna), tanmay@iitm.
ac.in (T. Basak), sjroy@iitm.ac.in (S. Roy), pop.ioan@yahoo.co.uk, popm.ioan@yahoo.
co.uk (I. Pop).
vection ow within lid driven porous cavities. Khanafer and Cham-
kha [15] reported mixed convection ow in a lid-driven enclosure
lled with a Darcian uid saturated porous medium in the pres-
ence of internal heat generation. Oztop [16] analyzed heat transfer
and uid ow in a partially heated porous lid driven cavity with
isothermal moving top wall. Khanafer and Vafai [17] analyzed
mixed convection heat and mass transport in a lid-driven square
cavity lled with a non-Darcian uid-saturated porous medium.
The two vertical walls of the enclosure are insulated, while the hor-
izontal walls are kept at constant but different temperatures and
concentrations with the top surface moving at a constant speed.
Kandaswamy et al. [18] investigated mixed convection heat trans-
fer in a lid-driven square cavity lled with a uid-saturated porous
medium. The left and right walls of the cavity are insulated,
whereas horizontal walls are kept at constant but different tem-
peratures. The top horizontal wall is moving in its own plane at a
constant speed, while all other walls are xed. Vishnuvardhanarao
and Das [19] investigated mixed convection heat transfer in a lid
driven square cavity lled with a uid-saturated porous medium
where left and right walls are moving upwards with same velocity.
The right wall is hot and left wall is maintained at constant cold
temperature. The top and bottom walls are xed and are thermally
insulated. However, most of the above numerical investigations
have been analyzed based on streamlines and isotherms and the
detailed analysis of heat ow was not well understood. Streamlines

0017-9310/$ - see front matter  2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2012.04.058
 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448
Nomenclature
Da Darcy number
A thermal aspect ratio
g acceleration due to gravity, m s2
k thermal conductivity, W m1 K1
H heatfunction
L length of the square cavity, m
K permeability, m2
Nu local Nusselt number
p pressure, Pa
P dimensionless pressure
Pr Prandtl number
Ra Rayleigh number
T temperature, K
Th temperature at the bottom edges of the side walls, K
Tc temperature at the top edges of the side walls, K
TH temperature at the center of the bottom wall, K
u x component of velocity
U x component of dimensionless velocity
v y component of velocity
V y component of dimensionless velocity
adequately depict uid ow whereas isotherms indicate only tem-
perature distribution which may not be useful as heat ow would
no more be in normal direction to isotherms for a convection dom-
inant regime.
The heatline technique is the best way to visualize the heat
transfer occurring in a two-dimensional convective transport pro-
cess. The technique was rst proposed by Kimura and Bejan [20]
and Bejan [21] which can be used to illustrate the path of heat ow,
its magnitude and zones of high heat transfer. Signicant number
of studies were carried out to visualize the heat ow for natural
convection in various applications [2226]. Zhao et al. [22] studied
double-diffusive convective ow of a binary mixture in a porous
enclosure subject to localized heating and salting from one side.
Varol et al. [23] used heatlines for natural convection heat transfer
in porous triangular enclosures with three different boundary con-
ditions (cases 13). Case 1 represents isothermal vertical and in-
clined walls whereas case 2 corresponds to adiabatic vertical
wall and isothermal inclined wall while case 3 involves isothermal
vertical wall and adiabatic inclined wall. In all three cases bottom
wall is non-isothermally heated. Hakyemez et al. [24] analyzed
heatlines to investigate the inuence on a heat barrier located in
the ceiling wall of an enclosure, on conjugate conduction/natu-
ral-convection heat transfer. Deng [25] investigated laminar natu-
ral convection in a two-dimensional square cavity due to two and
three discrete heat source-sink pairs on the vertical side walls. Da-
lal and Das [26] carried out natural convection inside a two-dimen-
sional cavity with a wavy right vertical wall whereas the bottom
wall is heated by a spatially varying temperature and other three
walls are kept at constant lower temperature. Recently, Kaluri
and Basak [27] analyzed the heat ow during natural convection
within discretely heated porous square cavities using heatlines
concept. They studied various cases depending on the location of
discrete heat sources on the walls of the cavity. However, a com-
prehensive analysis of inuence of thermal aspect ratio during
mixed convection in porous media based on heatlines is yet to ap-
pear in literature.
The current study deals with mixed convection within square
cavity lled with porous media for various thermal boundary con-
ditions based on thermal aspect ratio on bottom and side walls
where the top wall is adiabatic. An important non-dimensional
5437
X dimensionless distance along x coordinate
Y dimensionless distance along y coordinate
Greek symbols
a thermal diffusivity, m2 s1
b volume expansion coefcient, K1
c penalty parameter
h dimensionless temperature
m kinematic viscosity, m2 s1
q density, kg m3
U basis functions
w streamfunction
n horizontal coordinate in a unit square
g vertical coordinate in a unit square
Subscripts
b bottom wall
l left wall
r right wall
parameter thermal aspect ratio (A) is considered for thermal
boundary conditions. By varying thermal aspect ratio (A) from 0
to 1, various thermal boundary conditions are imposed on the bot-
tom and side walls, such as A = 0 corresponds to non-uniformly
heated bottom wall and cold side walls whereas A = 1 corresponds
to uniformly heated bottom wall and linearly heated side walls.
In the present study, numerical investigations on heating char-
acteristics within square cavities have been carried out based on
coupled partial differential equations of momentum and energy
which are solved using Galerkin nite element method with pen-
alty parameter to obtain the numerical simulations in terms of
streamfunctions, isotherms and heatfunctions. The streamfunc-
tions and heatfunctions of Poisson equation are also solved using
Galerkin nite element method. The jump discontinuity in Dirich-
let type of wall boundary conditions for temperature at the corner
points is tackled by implementation of exact boundary conditions
at those singular points. The heating and thermal mixing have been
studied for various lid velocities in terms of Re ranging from 1 to
100. Also, a number of model uid for almost all industrial applica-
tion ranges (0.015 6 Pr 6 7.2) have been considered for current
analysis.
Y, V
U=1,V=0 Adiabatic Wall
= A (1Y) = A (1Y)
g
= A + (1A) Sin ( X) X, U
Fig. 1. Schematic diagram of the physical system.
5438 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448

2. Mathematical formulation and simulation 107. Using Eq. (10), the momentum balance equations (Eqs. (2)
(4)) reduce to
2.1. Governing equations and boundary conditions   !
@U @U @ @U @V 1 @2U @2U
U V c
A schematic diagram of a two dimensional square cavity with @X @Y @X @X @Y Re @X 2 @Y 2
the physical dimensions is shown in Fig. 1. The top wall is assumed 1
to move from left to right with a uniform velocity U0. It is assumed  U; 11
Re Da
that the solid matrix of the porous media does not undergo defor-
and
mation and the porous bed is also assumed to be homogeneous,
isotropic, saturated with incompressible uid. All the thermophys-   !
@V @V @ @U @V 1 @2V @2V
ical properties of the uid such as viscosity, thermal conductivity, U V c
specic heats, thermal expansion coefcient and permeability ex-
@X @Y @Y @X @Y Re @X 2 @Y 2
cept the density variation in the buoyancy term are considered 1 Gr
to be constant. The Boussinesq approximation is considered for
 V 2 h: 12
Re Da Re
the body force term where the linear variation of density with tem-
The system of equations [Eqs. (4), (11) and (12)] with boundary
perature is assumed and that couples the temperature eld to the
conditions are solved by using Galerkin nite element method
ow eld. It is also assumed that the temperature of the uid phase
[28]. Since the solution procedure is explained in an earlier work
is equal to the temperature of the solid phase within the porous re-
[29,30], the detailed description is not included in this paper. The
gion, and local thermal equilibrium (LTE) is applicable in the pres-
numerical solutions are obtained in terms of velocity components
ent investigation [7]. Under these assumptions and neglecting the
(U, V).
Forchheimer inertia term, the governing equations for steady two-
dimensional mixed convection ow in a lid driven porous square
cavity using conservation of mass, momentum and energy can be 2.3. Streamfunction, Nusselt number and Heatfunction
written in dimensionless form as:
2.3.1. Streamfunction
@U @V The uid motion is displayed using the streamfunction (w) ob-
0; 1
@X @Y ! tained from velocity components U and V. The relationships be-
@U @U @P 1 @2U @2U 1 tween streamfunction, (w) and velocity components for two
U V   U; 2 dimensional ows are
@X @Y @X Re @X 2 @Y 2 Re Da
!
@V @V @P 1 @2V @2V 1 Gr @w @w
U V   V 2 h; 3 U and V  ; 13
@X @Y @Y Re @X 2 @Y 2 Re Da Re @Y @X
!
@h @h 1 @2h @2h which yield a single equation
U V : 4
@X @Y Re Pr @X 2 @Y 2
@2w @2w @U @V
2
 : 14
The relevant boundary conditions are @X @Y 2 @Y @X

U 0; V 0; h A 1  A sinpX 8Y 0; 0 6 X 6 1; 5 The no-slip condition is valid at all boundaries and there is no cross
U 0; V 0; h A1  Y; 8X 0; 0 6 Y 6 1; 6 ow. Eq. (14) has been solved by nite element method as dis-
cussed in earlier work [31,32]. The sign convention is as follows. Po-
U 0; V 0; h A1  Y; 8X 1; 0 6 Y 6 1; 7
sitive sign of w denotes anti-clockwise circulation and clockwise
@h circulation is represented by negative sign of w.
U 1; V 0; 0; 8Y 1; 0 6 X 6 1: 8
@Y
Here the dimensionless variables and parameters are dened as: 2.3.2. Nusselt number
The heat transfer coefcient in terms of the local Nusselt num-
x y u v T  Tc Th  Tc ber (Nu) is dened by
X ; Y ; U ; V ; h ; A
L L U0 U0 TH  Tc TH  Tc
p m U0 L gbT H  T c L3 K @h
P ; Pr ; Re ; Gr ; Da : 9 Nu  ; 15
2 a m m2 @n
qU 0 L2
where n denotes the normal direction on a plane. The local Nusselt
numbers at bottom wall (Nub), at the left wall (Nul) and at the right
2.2. Numerical solution procedure
wall (Nur) are dened as

The momentum and energy balance equations (Eqs. (2)(4)) X9

@ Ui
which are elliptic system of the equations have been solved using Nub hi ; 16
@Y
the Galerkin nite element method. The continuity equation (Eq. i1

(1)) is used as a constraint due to mass conservation and this con- X9

@ Ui
straint may be used to obtain the pressure distribution. In order to
Nul hi ; 17
i1
@X
solve Eqs. (2) and (3), we use the penalty nite element method
where the pressure, P, is eliminated by a penalty parameter c and
and the incompressibility criteria given by Eq. (1) results in
  X9
@ Ui
@U @V Nur  hi : 18
P c : 10 @X
@X @Y i1

The continuity equation (Eq. (1)) is automatically satised for large In general, the average Nusselt numbers at the bottom, left and
values of c. Typical values of c that yield consistent solutions are right walls in absence of any distributed heat sources are given by
 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448 5439

R1 Z 1
0
Nub dX 3. Results and discussion
Nub Nub dX; 19
Xj10 0
R1 Z 3.1. Numerical tests
1
0
Nul dY
Nul Nul dY; 20
Yj10 0 The computational domain consists of 28  28 bi-quadratic ele-
ments which correspond to 57  57 grid points. The bi-quadratic
and
elements with lesser number of nodes smoothly capture the non-
R1 Z
Nur dY 1 linear variations of the eld variables which are in contrast with -
0
Nur Nur dY: 21 nite difference or nite volume solution in the literature [33]. In the
Yj10 0
current investigation, Gaussian quadrature based nite element
method provides the smooth solutions at the interior domain
2.3.3. Heatfunction including the corner regions as evaluation of residuals depends on
The heat ow within the enclosure is displayed using the heat- the interior Gauss points and thus the effect of corner nodes are less
 @h @h

function (H) obtained from conductive heat uxes  @X ;  @Y as profound in the nal solution. The present nite element method
well as convective heat uxes (Uh, Vh). The steady energy balance based approach offers special advantage on evaluation of local Nus-
equations [Eq. (4)] can be arranged as selt number at the left, right and bottom walls as the element basis
      functions have been used here to evaluate the heat ux. The de-
@ 1 @h @ 1 @h
Uh  Vh  0: 22 tailed solution methodology is outlined in an earlier work [34].
@X Re Pr @X @Y Re Pr @Y
To assess the accuracy of the present formulation with non
The heatfunction satises the above equation such that Darcy Brinkman formulation, we have tested our algorithm based
on the grid size (57  57) of a porous square enclosure with iso-
@H 1 @h @H 1 @h
Uh  and  Vh  23 thermal cold and hot side walls and adiabatic horizontal walls
@Y RePr @X @X RePr @Y
and the results are in good agreement with the work of Lauriat
which yield a single equation and Prasad [35]. Results on mixed convection are benchmarked
with an earlier work and simulation results on isotherms and
@2H @2H @ @ streamlines are in good agreement with the earlier work [36].
Uh  Vh: 24
@X 2 @Y 2 @Y @X The detailed results are not shown here for the brevity of the man-
uscript. Table 1 shows that average Nusselt number Nub agrees
The sign convention for heatfunction is as follows. The positive sign with the results of earlier work [36] for a range of Grashof number
of H denotes anti-clockwise heat ow and clockwise heat ow is rep- (Gr = 104  105) and Prandtl number (Pr = 0.01, 0.1 and 1) at
resented by negative sign of H. Expanding the heatfunction (H) using Re = 500. The heatfunction for natural convection within square
P
the basis set U as H Nk1 Hk Uk X; Y. Eq. (24) can be solved using cavities are benchmarked with earlier results [37] and the results
Galerkin nite element method as discussed in earlier work [31,32]. were found to be in good agreement. The comparisons were shown
The boundary conditions of heatfunction are as follows. Neumann in an earlier work [31].
boundary conditions for H are obtained due to isothermal (hot or In the current investigation, numerical solutions on ow and
cold) wall based on Eq. (23) and they are specied as follows: temperature elds are obtained for visualization of heatlines with
various values of Da (Da = 105  103), Re (Re = 1  100), Pr
(a) for left wall (Pr = 0.015  7.2), Gr (Gr = 103  105) and A (A = 0.1  0.9) within
A the square cavity. The heat transfer in the cavity is conduction
n  rH 25 dominant for low values of the governing parameters (Re, Pr, Gr)
Re Pr
for any Da. Under these conditions, heatlines essentially represent
(b) for bottom wall heat ux lines, which are commonly used for conductive heat
p1  A cospX transport [31]. It is observed that heat ux lines emanate from
n  rH 26 hot surface and end on the cold surface and they are perpendicular
Re Pr
to the isothermal surfaces, similar to the heatlines, during conduc-
(c) for right wall
tion dominant regime. For A = 0.9, some heatlines are found to
A emanate from the hot portion of side wall and end on the relatively
n  rH  27
Re Pr cold portion of the side wall at higher Grashof number and Darcy
number (Gr = 105 and Da = 103) as seen from Fig. 4c. It is observed
and
that, the heat transfer within the cavity is conduction dominant for
(d) for top wall
Gr = 103 irrespective Da and Re for Pr = 0.015 (gures not shown). It
H constant adiabaticwall

The top insulated wall may be represented by Dirichlet bound- Table 1

Comparison of average Nusselt number for bottom wall with the benchmark
ary condition as obtained from Eq. (23) which is simplied into solutions of Moallemi and Jang [36] for mixed convection in a lid driven square
@H
@X
0 for an adiabatic wall. A reference value of H is assumed as cavity for various Grashof numbers (Gr) and Prandtl number (Pr) at Re = 500. Current
0 at X = 0,Y = 1 and hence H = 0 is valid for Y = 1, "X. It may be study employs 28  28 bi-quadratic elements (57  57 grid points).
noted that, the unique solution of Eq. (24) is strongly dependent Pr Gr Current work Moallemi and Jang [36]
on the non-homogeneous Dirichlet conditions. The following
Nu Nu
non-homogeneous Dirichlet boundary conditions are employed
to obtain the solution for Eq. (24). 0.01 104 1.0431 1.0167
0.01 105 1.0721 1.0368
1 0.1 104 2.3815 2.2382
H0; 0 Nul 0.1 105 2.8704 2.6290
RePr
1 1 104 5.5695 5.6089
H1; 0  Nur 28 1 105 6.3313 6.2118
RePr
5440 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448

is observed from gures that the uid circulations are strongly
dependent on Da. The detailed discussion on heat ow and uid
ow characteristics is carried out in following sections. The heat
transfer rates based on average Nusselt numbers are also evaluated
along bottom and side walls and they are explained based on
heatlines.
3.2. Effect of thermal aspect ratios on ow and thermal characteristics
Fig. 2 illustrates the streamlines, isotherms and heatlines for
Da = 105, Pr = 7.2 with Re = 1 at Gr = 105 for various A
(A = 0.1  0.9). Due to the drag force created by motion of upper
wall, a small quantity of uid is dragged to the top left corner of
the cavity. Two asymmetric cells with clockwise and anti-clock-
wise circulations are observed (Figs. 2ac). The magnitudes of
streamfunctions are very low and the isotherms are smooth and
symmetric illustrating conduction dominant heat transfer. At low
thermal aspect ratio (A = 0.1), the maximum value of streamfunc-
tion (jwjmax) is observed as 0.02 (Fig. 2a) and the isotherms except
h = 0.1 are smooth symmetric curves that span the entire cavity.
The isotherms with h = 0.1 also occur symmetrically near the side
walls of the enclosure. As thermal aspect ratio (A) increases to
0.5), the maximum value of streamfunction (jwjmax) is found to
be 0.013 (Fig. 2b). The isotherms with h 6 0.2 occur symmetrically
near the top portion of side walls of the enclosure, whereas h P 0.3
are smooth symmetric curves with respect to the vertical symmet-
ric line. At higher thermal aspect ratio (A = 0.9), the intensity of cir-
culation is smallest (jwjmax = 0.01) as compared to other thermal
aspect ratios (A = 0.1 and 0.5). The isotherms with h 6 0.3 are found
to be symmetric near the top portion of the side walls and the iso-
therms with h P 0.6 are almost parallel to bottom wall (Fig. 2c).
Heat ow distribution inside the cavity is illustrated by heat-
lines. Common to all thermal aspect ratios the heatlines are found
to be orthogonal to the isotherms signifying conduction dominant
heat transfer (Figs. 2ac). Heatlines are found to be dense near the
central zone of the bottom wall and hence the isotherms are dense
near that zone. Larger thermal gradient occurs near the center of
the bottom wall due to dense heatlines corresponding to
0.04 P H P 0.04 within 0.4 6 X 6 0.6 along the bottom wall
(Fig. 2a). As thermal aspect ratio (A) increases to 0.5, the thermal
gradient becomes smaller compared to that of A = 0.1 near the cen-
ter of the bottom wall due to the heatlines corresponding to
0.028 P H P 0.028 within 0.4 6 X 6 0.6 along the bottom wall
(see Fig. 2b). At higher thermal aspect ratio (A = 0.9), the heatlines
correspond to 0.016 P H P 0.016 occur within 0.4 6 X 6 0.6
along the bottom wall (Fig. 2c). The isotherms with h P 0.4 are lay-
ered along the bottom wall for A = 0.9, due to the vertical heatlines
based on conduction dominant heat transfer. It is found that, heat-
functions vary as 0.13 P jHj P 0.033,0.1 P jHj P 0.042 and
0.07 P jHj P 0.053 within 0 6 Y 6 0.5 along the side walls for
A = 0.1, 0.5 and 0.9, respectively (Fig. 2ac). Thus, the thermal
boundary layer thickness at lower portion of side walls is smaller
for A = 0.1 compared to that of A = 0.5 and 0.9. Note that, jHj varies
as 0.033 P jHj P 0, 0.042 P jHj P 0 and 0.053 P jHj P 0 within
0.5 6 Y 6 1 along the side walls for A = 0.1, 0.5 and 0.9, respectively
(Fig. 2ac). Consequently, the thermal boundary layer thickness at
upper portion of side walls is smaller for A = 0.9 compared to that
of A = 0.1 and 0.5.
Fig. 3 shows the streamlines, isotherms and heatlines for
Da = 104, Re = 1, Pr = 7.2 and Gr = 105 for various thermal aspect
ratios. At Da = 104, the streamlines become almost symmetric
due to the dominance of natural convection. The convection grad-
ually play a dominant role and the circulations near the central re-
gion become stronger. It may be noted that ow pattern is
qualitatively similar for A = 0.1, 0.5 and 0.9. At low thermal aspect
ratio (A = 0.1), the isotherms with h 6 0.3 start getting shifted par-
allelly towards the side walls illustrating dominant convection ef-
fect (Fig. 3a). The signicant convection is also indicated in other

(a) H
1 1 1 0.0
0.001 0 3
0.0 1 00
0.005 0.01 07 0. .01
0.1

0
0.1

0.01 0.2 0.0 02

0.013 2 0.
0.5 0.015 0. 0.5 0.5
017 0.0 4
0.4 4 0.0
0.02 0.07 7
0.02
0.7 0.0
0.1 . 1
0 0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(b)
1 1 1 0.00
0.007
0.1

1 05
0.1

0.003 0. 0.0 15
0.

0.0 2 0
0. 1
2

1 0. .0
0.007 0.3 0 0 03
0.011 0. 2 .
0.5 0.009 0.5 0.5 0 3 0
0.5 0.0
0.013 5 .05
0.012 0
0.7 0.0
7 .07
0.9 0
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(c)
1 1 1 0.
0 0.01 03

00
0.2

.0
0.

.3 1
0.

0
2

0
0.0 2
3

0.01 0.4
0.040.03 02
0.0
1

06
.

0.0
0.0 0.05 06

0.5 0.007 0.5 0.5

0.6
4

0.004 0.005
6

0.8
0.0

0.003
0.

0.003 0.9
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
5 5
Fig. 2. Streamfunction (w), temperature (h) and heatfunction (H) contours with Re = 1, Da = 10 , Pr = 7.2, Gr = 10 and (a) A = 0.1 (b) A = 0.5 (c) A = 0.9.
 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448 5441

(a) H
1 1 1
0.1 0.1 0.4 0.03 0.03
0.24 0.09 0.09
0.24

0.1
0.3
0.1
0.5

0.3
0.5 0.5 0.5
0.1
5 0.15
0 0.6
0.2
0.3 .3 0.2
0.8
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(b)
1 0.05 0.05 1 1
3
0.05 .0 0.05

0.
0

0.2
0.2

03
0.18 0.18 0.5
0.11

0.4
0.11

0.4
0.5 0.5 0.5 0.15 0.15
23

23
0.
0.

0.7
0.17 0.17
0.16 0.16
0.9
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(c)
1 1 1
0.03 0.03

0.3
0.09
0.3

0.09

0.5
0.5
0.6
0.5 0.16 0.16 0.5 0.5
0.7 2
0.1 0.12
0.1 0.1
0.05 0.05 0.9 0.07 0.07
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1

Fig. 3. Streamfunction (w), temperature (h) and heatfunction (H) contours with Re = 1, Da = 104, Pr = 7.2, Gr = 105 and (a) A = 0.1 (b) A = 0.5 (c) A = 0.9.

isotherms which are pulled towards the top portion. The maximum
value of streamfunction (jwjmax) is observed as 0.3 for A = 0.1. As
thermal aspect ratio increases to 0.5 (Fig. 3b), the isotherms with
h 6 0.4 occur towards the top portion of the side walls. The other
isotherms with h P 0.5 are pulled towards the top wall. The inten-
sity of streamfunction is found to be lesser compared to that of
A = 0.1 as jwjmax = 0.23. At higher thermal aspect ratio (A = 0.9),
the maximum value of streamfunction, jwjmax is observed as 0.16
(Fig. 3c). The isotherms with h 6 0.5 are symmetrically compressed
in a small region towards the top corners of the cavity at A = 0.9.
Interesting features of heat ow and temperature characteris-
tics are further explained based on heatlines. Some heatlines start-
ing from the bottom wall travel along central vertical line to reach
side walls and two symmetric convective circulations are formed
for all thermal aspect ratios (Figs. 3ac). At A = 0.1, the maximum
strength of heat ow near the core is observed as jHjmax = 0.2. Thus,
an isothermal zone with h = 0.4  0.5 is observed at the central top
portion of the cavity (Fig. 3a). On the other hand, the dense heat
ow near the core corresponding to jHjmax = 0.17 for A = 0.5. Thus,
uniform temperature is observed at the top central portion of the
cavity whereas h varies within 0.4  0.5 (see Fig. 3b). At A = 0.9,
the maximum intensity of heatline circulations (jHjmax) is observed
as 0.12 at the core which leads to the uniform temperature at the
top portion of the cavity (h = 0.5  0.6) (see Fig. 3c). It may be noted
that the temperature at the bottom portion is in general large at
higher A. Therefore, the top portion of the cavity reaches
h = 0.5  0.6 even with lower values of jHjmax at A = 0.9.
It is interesting to observe that, thermal gradient is large near
the center of the bottom wall for A = 0.1 and that decreases, as A in-
creases to 0.9. This is due to the heatlines corresponding to
0.14 P H P 0.14 for A = 0.1, 0.08 P H P 0.08 for A = 0.5 and
0.04 P H P 0.04 and for A = 0.9 within 0.2 6 X 6 0.8 along the
bottom wall (Fig. 3ac). It is observed that, jHj varies as
0.2 P jHj P 0.11, 0.12 P jHj P 0.08 and 0.07 P jHj P 0.06 for
A = 0.1, 0.5 and 0.9 respectively, within 0 6 Y 6 0.5 along the side
walls. Therefore, the thermal boundary layer thickness is smaller
for A = 0.1 along the lower portion of side walls. Also, jHj varies as
0.11 P jHj P 0, 0.08 P jHj P 0 and 0.06 P jHj P 0 for A = 0.1, 0.5
and 0.9 respectively, within 0.5 6 Y 6 1 along the side walls. Thus
the thermal boundary layer thickness is smaller for A = 0.1 along
the upper portion of side walls.
Fig. 4 shows the streamlines, isotherms and heatlines for
Da = 103, Re = 1, Pr = 7.2 and Gr = 105 for various thermal aspect
ratios. It may be noted that the circulation near the core becomes
stronger irrespective of A as seen from Figs. 4ac. The isotherms
with h 6 0.4 for A = 0.1 and h 6 0.5 for A = 0.5 and 0.9 are com-
pressed towards the top portion of side walls (Figs. 4ac). It may
be noted that, the intensity of circulations decreases with thermal
aspect ratio, as jwjmax is represented with 1.5, 1.45 and 1 for A = 0.1,
0.5 and 0.9, respectively (Figs. 4ac). It is interesting to observe
that weak symmetric secondary circulations appear at the bottom
corners of the cavity for A = 0.9 (Fig. 4c).
The effect of enhanced convection can be explained effectively
based on heatlines. The heatline cells are found to be qualitatively
similar to streamlines which indicate convection dominant heat
transfer. At low A (A = 0.1), the maximum intensity of heat ow
may be observed as jHjmax = 0.75 at the core (see Fig. 4a). Thus, a
large central portion is maintained with h = 0.4  0.5 at the upper
regime of the cavity. On the other hand, the heat ow at the core
is represented with jHjmax = 0.85 for A = 0.5 and thus, a large iso-
thermal zone with h = 0.5  0.6 is found at the upper portion of
the cavity (see Fig. 4b). At higher thermal aspect ratio (A = 0.9),
the secondary heatline circulations push the primary circulations
towards the upper portion of the cavity and jHjmax is found to be
0.6. Uniform temperature is observed with h = 0.5  0.6 at the
upper portion of the enclosure for A = 0.9 (Fig. 4c).
It is found that, the isotherms with h P 0.6 for A = 0.1 and h P 0.7
for A = 0.5 are highly compressed along the bottom wall, whereas
5442 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448

(a) H
1 1 1
0.1 0.1 0.

.1
0.7 1

0
0 .4
0.5

1 0.7

.4
1.

0.4

0.
3

.3

5
5
4

0.
0.5 0.5 0.5

0.
0.
75 .75
1.5 1.5 0.3 0.3 0
0.6

0.

2
2

0.
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(b)
1 1 1 0. 05
0.3 .3 0.7 0.3 05 0. 0.3
0.7 1 0

0
0.5
0.6

.5
1

0.

.6
5

0
0.3
0.
0.5 0.5 0.5

0.3
1.

0.
45 0.8

5
45

5
1.

.8
5

0
4
0.

0.
0.7

4
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(c)
1 1 1
0.1
0.7 0.3

.1
0.3
.3

0.4

0
0.3
7 0.6 .4
0

0. 0.
5 0
5
0. 0.6
0.5 0.4 0.4 .6
0.5 1 1 0.5 0
7
0.
0.4 0.4 0.1 0.1
0 0 0.9 0
0 0.5 1 0 0.5 1 0 0.5 1

Fig. 4. Streamfunction (w), temperature (h) and heatfunction (H) contours with Re = 1, Da = 103, Pr = 7.2, Gr = 105 and (a) A = 0.1 (b) A = 0.5 (c) A = 0.9.

the isotherms with h P 0.7 are less compressed for A = 0.9. This is
due to the dense heatlines corresponding to 0.35 P H P 0.35 for
A = 0.1, 0.23 P H P 0.24 for A = 0.5 and 0.17 P H P 0.17 for
A = 0.9 within 0.2 6 X 6 0.8 along the bottom wall. It is observed
that heatfunctions vary as 0.47 P jHj P 0.35, 0.32 P jHj P 0.31
and 0.22 P jHj P 0.19 within 0 6 Y 6 0.5 along the side walls for
A = 0.1,0.5 and 0.9, respectively. Thus, the boundary layer thickness
at lower portion of the side walls is lesser for A = 0.1 compared to
A = 0.5 and 0.9 (see Figs. 4ac). It may also be noted that, heatfunc-
tions vary as 0.35 P jHj P 0, 0.31 P jHj P 0 and 0.19 P jHj P 0
within 0.5 6 Y 6 1 along the side walls for A = 0.1, 0.5 and 0.9,
respectively. Therefore, the thermal boundary layer thickness is les-
ser for A = 0.1 compared to A = 0.5 and 0.9 along the upper portion of
side walls.
At Da = 105 with Pr = 0.7 and Re = 10, the combined inuence
of buoyancy and lid-driven force is observed inside the cavity (Fig-
ure not shown). Due to the drag force created by the upper lid
asymmetric pattern is found in the left and right circulation cells.
The right circulation is dominant due to the movement of upper
lid. The magnitudes of streamfunctions are very low signifying
conduction dominant heat transfer. The smooth and symmetric
isotherms also represent the conduction dominant heat transfer
irrespective of A. The maximum intensity of left circulation cells
is represented with 0.0015, 0.0007 and 0.0001 for A = 0.1, 0.5 and
0.9, respectively with very weak uid ow being observed at the
right half of the cavity (Figure not shown). The isotherms and heat-
lines are qualitatively similar to that of Da = 105,Pr = 7.2, with
Re = 1 and similar explanation as that of Fig. 2 follows.
Fig. 5 displays the distribution of streamfunction (w), tempera-
ture (h) and heatfunction (H) contours for Re = 10, Da = 104,
Pr = 0.7, Gr = 105 for various thermal aspect ratios (A = 0.1  0.9).
As Da increases to 104, circulations near left side become stronger
due to gradual dominance of natural convection. The maximum
value of primary circulations (jwjmax) at the right half of the cavity
is found to be 0.019, 0.012 and 0.012 for A = 0.1, 0.5 and 0.9,
respectively whereas that is represented with 0.018, 0.011 and
0.0045 for secondary circulations at the left portion of the cavity
for A = 0.1, 0.5 and 0.9, respectively. The temperature distribution
and heat energy distributions are almost qualitatively similar to
previous case and Fig. 2.
It is observed that the distribution in streamlines, isotherms
and heatlines for Re = 10 with Pr = 0.7 and Da = 103 are qualita-
tively similar to that of Re = 1 with Pr = 7.2 and Da = 104 and qual-
itatively similar explanation as that of Fig. 3 follows for Re = 10
with Pr = 0.7 at Da = 103 (Figure not shown). It is also found that
streamlines, isotherms and heatlines for Re = 10 are qualitatively
similar to that of Re = 1 at Pr = 7.2 with Da = 103 (Fig. 4). It is inter-
esting to observe that, at Re = 100 with Da = 105, Pr = 0.7 and
Gr = 105 the primary circulation cells occupy most of the cavity ex-
cept a small region consisting of secondary cells which occupy at
the bottom left corner of the cavity for A = 0.1 (gure not shown).
The isotherms with h 6 0.1 are found to be asymmetric and com-
pressed towards the right wall for A = 0.1. As A increases to 0.5,
the secondary circulations tend to vanish and primary circulations
occupy most of the cavity. The isotherms with h 6 0.2 are stretched
and asymmetrically shifted towards the top corners of side walls.
At higher aspect ratio (A = 0.9), the primary circulations occupy
the entire cavity. The isotherms with h 6 0.3 are stretched and
found to be compressed along the top corners of side walls. It is
interesting to observe that heatline distribution for Re = 100 with
Da = 105 is qualitatively similar to that of Re = 10 with Da = 104
at Pr = 0.7 (see Fig. 5).
Fig. 6 illustrates the streamlines, isotherms and heatlines for
Re = 100, Da = 104, Pr = 0.7, Gr = 105 for various thermal aspect
ratios (A = 0.1  0.9). At A = 0.1, maximum intensity of the left cir-
culation cell is found to be 0.001, whereas that is observed as 0.007
for right circulation cells. The isotherm with h = 0.1 occurs asym-
metrically towards the side walls. At A = 0.5 the intensity of
 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448 5443

(a) H
1 1 1

0.001 05
05
0.008
0.

0.1

0.1

.0
0 .0
00
0.013 0.2

0
15
0.01 3
0.5 0.5 0.5 02

0.
.
0

02
0.4 0.0 .05

3
0.018 0.019 0.6
5 0
0.8 0.1 0.1
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(b)
1 1 1

0.1
0.1
5

0.001
.01
0
2

0.
0. 0
0.012
0.0
0.008

2
0.3

0.
.00
05

0.01

01
35
15

5
0.5 0.5 0.4 0.5 .0

0.
0

03
5
0.01 0.6
1 0.
0.8 07 .07
0
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(c)
1 1 1
0

0.
2

0.005
0.
0.
0.

.3 3
.00

2
012 0 0.4
8

0.005
45

0.02
0.0

0.5
0.00

0.5 0.5 0.5

0.02

4
0.6

0.04

0.0
015

6
0.003 0.8

0.06

0.0
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1

Fig. 5. Streamfunction (w), temperature (h) and heatfunction (H) contours with Re = 10, Da = 104, Pr = 0.7, Gr = 105 and (a) A = 0.1 (b) A = 0.5 (c) A = 0.9.

(a) H
1 1 1
0.007
015
0.

0.0
1

0.001
1
0.

0.00 2

0.004 0.2
0.5 0.5 0.5 003
0.
0.
0

01
00

0.4
.0

0.001 0.
01

00
0.002 5 007
0.7 0.0
1 0. .01
0 0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(b)
1 1 1
1
0.1

0. 0.
0

0.007 0.
0.002

00
0.2 2 01
.0
01

0.3
0.004 03
.0
0.

0.5 0.5 0.4 0.5 0

00

0
2

.0 0.0 5
01 0.002
0.00
05 0.6 05 .00
0
0.8 0.008
0 0 0 0.008
0 0.5 1 0 0.5 1 0 0.5 1
(c)
1 1 1
0
0.2
0.3

0.2 0.3
.0
03
0. 01

0.005 0.4
0.0

04
001
0

0.5 0.5 0.5 0.5

0.0
5

0.002
0
0.00

0.7
.001

1e
0.00

0.00

0 0.000
3

05 5
5

0.9
5

0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
4 5
Fig. 6. Streamfunction (w), temperature (h) and heatfunction (H) contours with Re = 100, Da = 10 , Pr = 0.7, Gr = 10 and (a) A = 0.1 (b) A = 0.5 (c) A = 0.9.

secondary circulations gradually become weaker. The maximum with h 6 0.2 are clustered towards to the top portion of side walls
strength of the left circulation cell is observed as 0.0005 whereas and the isotherms are found to be asymmetric. At higher A
that is found to be 0.007 for right circulation cells. The isotherms (A = 0.9), very less intense secondary circulation cells are observed
5444 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448

at bottom left corner. The isotherms are found to be asymmetric
and h P 0.4 are slightly pulled towards the top wall along the ver-
tical central line with inclination towards the left wall.
Heat ow distribution inside the cavity is illustrated by heatlines.
It is observed that, dense heatlines appear with 0.0045 P H P
0.0037 for A = 0.1, whereas those occur with 0.0031 P H P
0.0024 for A = 0.5 and 0.0017 P jHj P  0.0012 for A = 0.9 within
0.4 6 X 6 0.6 along the bottom wall. Consequently, thermal bound-
ary layer thickness near the central portion of bottom wall is less for
A = 0.1 and that increases as A increases to 0.9 (Figs. 6ac). It may be
noted that, the heatfunctions vary as 0.0133 P jHj P 0.0032,
0.0097 P jHj P 0.004 and 0.0062 P jHj P 0.0049 within 0 6 Y 6
0.5 along the side wall wall for A = 0.1,0.5 and 0.9, respectively (Figs.
6ac). Thus, thermal gradient is higher for A = 0.1 compared to that of
A = 0.5 and 0.9 along the bottom portion of side walls. It is also ob-
served that, heatfunctions vary as 0.0032 P jHj P 0, 0.004 P jHj P
0 and 0.0049 P jHj P 0 for A = 0.1, 0.5 and 0.9, respectively within
0.5 6 Y 6 1 along the side walls. Thus, the thermal gradient is larger
for A = 0.9 compared to that of A = 0.1 and 0.5 along the top portion of
the side walls (Figs. 6ac).
Fig. 7 displays the distribution of streamfunction (w), tempera-
ture (h) and heatfunction (H) contours for Re = 100, Da = 103,Pr =
0.7 and Gr = 105 for various thermal aspect ratios (A = 0.1  0.9).
The signicant natural convection is clearly observed at higher
Darcy number (Da = 103) as the uid rises from the core and is
pulled towards the top left corner of the cavity especially for
A = 0.1 and 0.5 as seen in Fig. 7a and b. The maximum intensity of
streamfunction for left circulation cells is observed as jwjmax = 0.015,
0.01 and 0.001 for A = 0.1, 0.5 and 0.9, respectively (Fig. 7ac). On
the other hand, the maximum value of streamfunction for right cir-
culation cells is represented with 0.032, 0.03 and 0.027 for A = 0.1,
0.5 and 0.9, respectively. The isotherms with h 6 0.2 are clustered
asymmetrically towards the side walls for A = 0.1, whereas the iso-
therms with h 6 0.2, and h 6 0.3 are clustered towards the top por-
tion of side walls for A = 0.5 and 0.9, respectively.
Some interesting features in ow pattern and temperature dis-
tributions are further explained based on heatlines. It is interesting
to observe that dense heatlines are found along an inclined path and
isotherms are distorted within the cavity for all thermal aspect ra-
tios. At A = 0.1, isotherms with h P 0.6 are highly compressed along
the bottom wall due to dense heatlines with 0.004 P H P  0.0045
within 0.5 6 X 6 0.7 (Fig. 7a). As thermal aspect ratio (A) increases
to 0.5, the isotherms with h P 0.7 are less compressed along the
bottom wall compared to that of A = 0.1. It is observed that heat-
function, H varies as 0.0046 P H P  0.0043 within 0.5 6 X 6 0.8
along the bottom wall (Fig. 7b). At higher thermal aspect ratio
(A = 0.9), thermal gradient is less compared to A = 0.1 and 0.5. It is
observed that heatfunctions vary as 0.005 P H P  0.000047 with-
in 0.5 6 X 6 0.8 along the bottom wall. It is also found that heatfunc-
tions vary as 0.0193 P jHj P 0.0087 for A = 0.1, 0.013 P jHj P
0.0079 for A = 0.5 and 0.01 P jHj P 0.008 for A = 0.9 within
0 6 Y 6 0.5 along the left wall (Figs. 7ac). Thus, the thickness of
thermal boundary layer is smaller for A = 0.1 compared to that of
A = 0.5 and A = 0.9 along the bottom portion of left wall. Note that,
jHj varies as 0.0087 P jHj P 0, 0.0079 P jHj P 0 and 0.008 P jHj P
0 within 0.5 6 Y 6 1 for A = 0.1, 0.5 and 0.9, respectively along the
top portion of the left wall. Thus, the thermal boundary layer is less
for A = 0.1 compared to that of A = 0.5 and A = 0.9 along the top por-
tion of left wall. It is also found that, heatlines corresponding to
0.015 P jHj P 0.0078, 0.0078 P jHj P 0.0059 and 0.0043 P jHj P
0.0023 occur within 0 6 Y 6 0.5 along the right wall for A = 0.1,0.5
and 0.9, respectively (Figs. 7ac). Thus, the thermal boundary layer
thickness is smaller for A = 0.1 along the bottom portion of right
wall. Also, jHj varies as 0.0078 P jHj P 0, 0.0059 P jHj P 0 and
0.0052 P jHj P 0 within 0.5 6 Y 6 1 along the right wall for
A = 0.1, 0.5 and 0.9, respectively. As a result, thermal boundary layer
thickness is small for A = 0.1 along the top portion of right wall as
seen from Figs. 7ac.
Fig. 8 illustrates the streamlines, isotherms and heatlines for
Re = 100, Da = 103, Pr = 7.2, Gr = 105 for various thermal aspect

(a) H
1 1 1
07
0

3 0.0
0.0 0.00
0.2
2

.0

0.3 2
0.

0.001
03

0.005
0.1

0.1

0.4 1
.01 .0
0.5 0 0.5 0.5 0.01 0
0

0.032
14
.0

0.0
12

0.6
5 0.01
0.01 18 4
0.015 0.8 0.0
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(b)
1 1 1
0.3 0.008
0.
2

0.
0.

0
0.03 0.4 0.005 01 0.01
0.2

0.001
005

0.5
0.5 0.007 0.027 0.5 0.5
0.01
0.012
0.02 0.7
0.01 4
0.005 0.9 0.01
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(c)
1 1 1
0.2

0.3
0. 005

0.4
0.3

0.0.00 0.00

0.0
0

005

27 0.5
4

0.5 0.02 0.5 0.5 0.01

0.7
0.007
7

0.01
0.01

0.001 0.9
0 0.001 0 0
0 0.5 1 0 0.5 1 0 0.5 1
3 5
Fig. 7. Streamfunction (w), temperature (h) and heatfunction (H) contours with Re = 100, Da = 10 , Pr = 0.7, Gr = 10 and (a) A = 0.1 (b) A = 0.5 (c) A = 0.9.
 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448 5445

ratios (A = 0.1  0.9). The dominant lid driven force is observed at
A = 0.1 as a single primary circulation spans almost entire cavity
except at bottom left corner of the cavity consisting of a secondary
circulation cell. The maximum strength (jwjmax) of the left circula-
tion cells is 0.001 whereas that is found to be 0.027 for right circu-
lation cells. The isotherms with h P 0.2 are highly compressed
towards the left wall and a large portion near the right wall is
maintained with h 6 0.2. As A increases to 0.5, multiple circulation
cells are observed with jwjmax = 0.023, 0.002 and 0.002 at upper,
bottom left and central right portions of the cavity, respectively.
The isotherms are distorted and h 6 0.1 occupy almost 75% of the
cavity. At higher A (A = 0.9), multiple circulations in uid ow are
found with jwjmax = 0.023, 0.001, 0.004, 0.001 and 0.005 at various
locations of the cavity. Highly distorted isotherms are observed
and h P 0.2 occupy almost 70% of the cavity.
Many interesting features of heatlines are observed for Re = 100,
Da = 103, Pr = 7.2, Gr = 105 for various thermal aspect ratios
(A = 0.1  0.9). A large portion is maintained with uniform temper-
ature as h varies within 0.10.2 for all thermal aspect ratios. It may
be noted that the isothermal zone with h = 0.1  0.2 occupies almost
50% of the cavity for A = 0.1, whereas that occupies almost 25% for
A = 0.5 and 30% for A = 0.9 within the cavity. This is due to the larger
intensity of heatfunctions corresponding to jHjmax = 0.0048 for
A = 0.1, whereas that is observed as jHjmax = 0.002 for A = 0.5 and
jHjmax = 0.003 for A = 0.9 within upper portion of the cavity. At low
thermal aspect ratio (A = 0.1), isotherms with h P 0.2 are highly
compressed with larger thermal gradient along the bottom wall
due to highly dense heatlines corresponding to 0.0036 P jHj P
0.0038 within 0.4 6 X 6 0.7 along the bottom wall (Fig. 8a). As A in-
creases to 0.5, isotherms are less compressed along the bottom wall
compared to that of A = 0.1. It is observed that heatlines occur with
0.00087 P H P  0.00016 within 0.5 6 X 6 0.8 along the bottom
wall. At higher A (A = 0.9), thermal gradient is low along the bottom
wall due to sparse heatlines at that zone.
It may be noted that the heatfunctions vary as 0.0044 P jHj P
0.0023, 0.0017 P jHj P 0.0013 and 0.0018 P jHj P 0.0012 within
0 6 Y 6 0.5 along the left wall for A = 0.1, 0.5 and 0.9, respectively
(Figs. 8ac). Thus, thermal boundary layer thickness is smaller for
A = 0.1 compared to that of A = 0.5 and 0.9 along the bottom portion
of left wall. It is also observed that, heatfunctions vary as
0.0023 P jHj P 0, 0.0013 P jHj P 0 and 0.0018 P jHj P 0 within
0.5 6 Y 6 1 along the left wall, for A = 0.1,0.5 and 0.9, respectively.
Thus, the thermal boundary layer thickness is smaller for A = 0.1
compared to that of A = 0.5 and 0.9 along the top portion of the left
wall (Figs. 8ac). It is also observed that, jHj values vary as
0.0018 P jHj P 0.0016, 0.0003 P jHj P 0.00024 and 0.00067 P jHj
P 0.00053 within 0 6 Y 6 0.5 along the right wall for A = 0.1,0.5 and
0.9, respectively (Figs. 8ac). Thus, the thermal boundary layer
thickness is also found to be smaller for A = 0.1 along the bottom
portion of right wall. Also, jHj varies as 0.0016 P jHj P 0,
0.00024 P jHj P 0 and 0.00053 P jHj P 0 within 0.5 6 Y 6 1 along
the right wall for A = 0.1, 0.5 and 0.9, respectively. As a result, ther-
mal boundary layer thickness is small for A = 0.1 along the top por-
tion of right wall as seen from Figs. 8ac.
3.3. Overall Heat transfer rates and average Nusselt numbers
Figs. 9ab illustrate the overall heat transfer rates for the bottom,
left and right walls for Re = 1  100,Da = 105  103, Pr = 0.7  7.2
with various thermal aspect ratios (A = 0.1  0.9). Fig. 9a displays
the average Nusselt number distribution for bottom wall at Re = 1.
At low Darcy number (Da = 105), Nub distribution is constant for
the entire range of Gr for all thermal aspect ratios (A) and Prs (see
the lower panel of Fig. 9a). The constant values are clearly observed
from the range of heatfunction values at Gr = 103 and Gr = 105. Note
that, heatlines corresponding to 1.34 P jHj P 0.07 for A = 0.1,
1.04 P jHj P 0.05 for A = 0.5 and 0.74 P jHj P 0.02 for A = 0.9 occur
at Gr = 103, whereas those appear with 1.35 P jHj P 0.07 for A = 0.1,

(a) H
1 1 1
0.0042
0.0

27 48
0.0 00
0.2

01
0.1

.
0
0.3

0.5 2 0.5 0.5

.0
0.002

0
03
0.

.0
01

0
4
.
0.0

0
01

0.6
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(b)
1 1 1
0.0 0.0004

0.023
0.002
015

01
0. 0.1 08
.00
0.2

0.5 0.5 0.5 04

0.002 0 0.00
0.3
0.00 0.5 0.002 0.0008
2 0.7
0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
(c)
1 1 1
0.0 0.0001

0.023 0.003
0.2 0.3

02

01
0.0
1 0. 0.0
0.5 00 0.5 0.5
0.

0.0 4 0.001
0

01
02

0.4
0.003

0.0 0.5
05

0.00
01 0.8 0.7 2
.0
0

0 0 0
0 0.5 1 0 0.5 1 0 0.5 1
3 5
Fig. 8. Streamfunction (w), temperature (h) and heatfunction (H) contours with Re = 100, Da = 10 , Pr = 7.2, Gr = 10 and (a) A = 0.1 (b) A = 0.5 (c) A = 0.9.
5446 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448

1.04 P jHj P 0.05 for A = 0.5 and 0.74 P jHj P 0.02 for A = 0.9 at
Gr = 105 (Da = 105, Pr = 0.7 and Re = 1). It may also be noted that
Nub at Pr = 0.7 and Pr = 7.2 are almost identical at Da = 105 with
Re = 1 (see the lower panel of Fig. 9a).
The upper panel of Fig. 9a illustrates the Nub distribution for
Da = 103 with Re = 1. At higher Darcy number (Da = 103) with
Pr 0:7; Nub is almost constant within 103 6 jHj 6 6  104 for
A = 0.1, thereafter that increases slightly upto Gr = 105 (see the upper
panel of Fig. 9a). This is due to the larger gradients of heatfunctions
1.80 P jHj P 0.002 at Gr = 105, whereas those are observed as
1.53 P jHj P 0.002 at Gr = 6  104 and 1.35 P jHj P 0.018 at
Gr = 103. Note that, Nub remains almost constant for A = 0.5 and 0.9
(see the upper panel of Fig. 9a). It is found that, jHj varies as
1.04 P jHj P 0.002 for A = 0.5, 0.74 P jHj P 0.002 for A = 0.9 at
Gr = 103, whereas those are observed as 1.1 P jHj P 0.002 for
A = 0.5, 0.7 P jHj P 0.002 for A = 0.9 at Gr = 105 (Da = 103,
Pr = 0.7). It is observed that Nub at A = 0.1 with Pr = 7.2 is almost con-
stant within 103 6 Gr 6 5  103, thereafter that increases smoothly
upto Gr = 105 (see the upper panel of Fig. 9a). It may be noted that
jHj varies as 0.13 P jHj P 0.001 at Gr = 103, whereas that varies as
0.14 P jHj P 0.002 at Gr = 5  103 and 0.47 P jHj P 0.001 at
Gr = 105 for A = 0.1 (Da = 103,Pr = 7.2). Qualitatively similar distri-
bution is observed in Nub distribution for A = 0.5 and A = 0.9 com-
pared to that of A = 0.1 (see the upper panel of Fig. 9a).
Fig. 9b displays the distribution of average Nusselt number at left
wall and the inset plot displays the distribution of average Nusselt
number at the right wall. It is interesting to observe that Nul and
Nur distributions are qualitatively similar to that of Nub
distribution irrespective of A, Pr and Da (see Fig. 9a and b). Fig. 9c
and d illustrate the average Nusselt number for bottom, left and right
walls at Re = 100 with A = 0.1  0.9, Pr = 0.7  7.2 and Da =
105  103. It is observed that Nub distribution at Re = 100 is
qualitatively similar to that of Re = 1 irrespective of Da, Pr and A, ex-
cept at A = 0.5 and 0.9 with Pr = 7.2, Da = 103 (see the upper and
lower panels of Fig. 9a and c). Note that, Nub at Da = 103 with
Pr = 7.2 increases within 103 6 Gr 6 2  104, thereafter that de-
creases upto Gr = 105 for A = 0.5. It may be noted that jHj at A = 0.5
varies as 0.002 P jHj P 0 at Gr = 103, whereas that varies as
0.0025 P jHj P 0 at Gr = 2  104 and 0.0017 P jHj P 0 at Gr = 105
(Da = 103, Pr = 7.2). Fig. 9d displays the average Nusselt number
for left wall and the inset plot illustrates the distributions for right
wall. It is interesting to observe that Nul and Nur distributions are
qualitatively similar to that of Nub distribution irrespective of A,Pr
and Da except at A = 0.5 and 0.9 with Da = 103 with Pr = 7.2 (see
Fig. 9c and d). Note that, Nul distribution for A = 0.5 is almost con-
stant within 103 6 Gr 6 104, thereafter that decreases upto Gr = 105
for Da = 103 with Pr = 7.2 (see inset plot in upper panel of Fig. 9d).
Qualitatively similar distribution is observed in Nul distribution for

(a) Bottom Wall (b) Left Wall

8
Right Wall
Da=10-3 A=0.9 Da=10-3
6 A=0.9
A=0.5
6 A=0.1 A=0.5
2
Nur

4 A=0.9
A=0.1
Average Nusselt Number, Nub

Average Nusselt Number, Nul

4 A=0.5
0
2 2 A=0.1

0 0
Da=10-5 A=0.1 Right Wall Da=10-5
3 A=0.9 A=0.5 A=0.1
2 1
A=0.1
Nur

A=0.5
2 A=0.5
A=0.9 0
A=0.9
1 1

0 3
103 104 105 10 104 105
Grashof Number, Gr Grashof Number, Gr

(c) Bottom Wall (d) Left Wall

6 Right Wall
Da=10-3 8 2
A=0.1 Da=10-3
A=0.9 A=0.5 A=0.1
A=0.9 A=0.5 A=0.1
Nur

4 6 1
A=0.5
Average Nusselt Number, Nub

Average Nusselt Number, Nul

4 0
2 A=0.9
2

0
0 Right Wall
Da=10-5 A=0.5 A=0.1 Da=10-3
A=0.9 A=0.5 A=0.1
A=0.9 1
Nur

2 A=0.1
2
0
A=0.9 A=0.5
1

0 3
103 104 105 10 104 105
Grashof Number, Gr Grashof Number, Gr

Fig. 9. Variation of average Nusselt number with Grashof number (Gr) with Re = 1 [a and b] and Re = 100 [c and d], Gr = 105, Pr = 0.7 (-) and Pr = 7.2 (- - -). In each plot, upper
panel corresponds to Da = 103 and lower panel corresponds to Da = 105.
 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448 5447

(a) Re = 1 (b) Re = 100

9
6
6

Nub

Nub
3
3

0 0
Nul 4 4

Nul
2 2

0 0
4 2

Nur
Nur

2 1

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Thermal aspect ratio, A Thermal aspect ratio, A

Fig. 10. Variation of average Nusselt number with thermal aspect ratio for (a) Re = 1 and (b) Re = 100 with Da = 103 and Gr = 105; Pr = 0.7 (-) and Pr = 7.2(. . . .).

A = 0.5 and A = 0.9 at Da = 103 with Pr = 7.2 (see upper panel of
Fig. 9d). It may be noted that, Nur distribution for A = 0.5 is almost
constant within 103 6 Gr 6 104, thereafter that decreases upto
Gr = 5  104 for Da = 103 with Pr = 7.2. Further Nur increases upto
Gr = 105 (see inset plot in upper panel of Fig. 9d). Note that, Nur dis-
tribution at A = 0.9 is qualitatively similar to that of A = 0.5 at
Da = 103 with Pr = 7.2 (see inset plot in upper panel of Fig. 9d).
Fig. 10a and b illustrate the distributions of the average Nusselt
number at bottom and side walls vs. thermal aspect ratio (A =
0  1) for Pr = 0.7  7.2, Da = 103 and Gr = 105 with Re = 1 and 100
respectively. It is found that Nub , Nul and Nur decreases as A increases
from 0 to 1, irrespective of Pr and Re except at Re = 100 with Pr = 7.2
(see Fig. 10a and b). It is observed that Nub at Re = 100 with Pr = 7.2
decreases within 0 6 A 6 0.55, thereafter that increases upto A = 0.8
and that decreases slowly within 0.8 6 A 6 0.94. Further, Nub in-
creases upto A = 1 (see upper panel of Fig. 10b). It is observed that
heatfunctions vary as 0.0049 P jHj P 0 at A = 0, whereas those are
vary as 0.0013 P jHj P 0 at A = 0.55, 0.00175 P jHj P 0 at A = 0.8,
0.0016 P jHj P 0 at A = 0.94 and 0.0029 P jHj P 0 at A = 1 along
the bottom wall (Re = 100,Pr = 7.2). It is interesting to observe that
Nul distribution is qualitatively similar to that of Nub distribution
at Re = 100 with Pr = 7.2 (see upper and middle panels of Fig. 10b).
Note that Nur distribution at Re = 100 with Pr = 7.2 decreases within
0 6 A 6 0.45, thereafter that increases upto A = 0.6. Further, Nur re-
mains almost constant within 0.6 6 A 6 1. It is found that, jHj varies
as 0.0048 P jHj P 0 at A = 0, whereas those vary as
0.0021 P jHj P 0 at A = 0.45, 0.0025 P jHj P 0 at A = 0.6,
0.0027 P jHj P 0 at A = 1 (Re = 100, Pr = 7.2).
4. Conclusion
of streamlines, isotherms and heatlines for various Prandtl num-
bers Pr (Pr = 0.015  7.2), Da (Da = 105  103), Re (Re = 1  100)
and Gr (Gr = 103  105), with thermal aspect ratio (A = 0.1  0.9).
Important features on heating ow characteristics are outlined
for various studies.
 At low Reynolds number (Re = 1), almost symmetric prole is
observed in isotherms and heatlines. It is found that the effect
of lid driven ow is dominant for Da = 105, whereas buoyancy
force is dominant for Da = 104 and Da = 103. Secondary circu-
lations are found at the bottom corners of the cavity for higher
thermal aspect ratio (A = 0.9) with Da = 103 and Pr = 7.2.
 At high Reynolds number (Re = 100), asymmetric prole is
observed in temperature and heatfunction distribution. Multi-
ple circulations are found in streamlines and heatlines espe-
cially for A = 0.5 and 0.9 at high Reynolds number (Re = 100)
with Da = 103 and Pr = 7.2.
 Larger thermal gradients are observed at the center of the bot-
tom wall at A = 0.1 due to dense heatlines at that zone and that
decreases as A increases from 0.1 to 0.9, irrespective of Re,Da
and Pr.
 The average Nusselt number show that the overall heat transfer
rate is high at A = 0.1 compared to that of A = 0.5 and A = 0.9
irrespective of Reynolds number (Re) and Darcy number (Da).
 The variation of average Nusselt number with thermal aspect
ratio illustrate that, overall heat transfer rates decrease as ther-
mal aspect ratio (A) increases from 0 to 1, for Re = 1 irrespective
of Pr, whereas that exhibits oscillations for higher thermal
aspect ratios at Pr = 7.2 with Re = 100.

The effect of various thermal boundary conditions based on Acknowledgement

thermal aspect ratio (A) on the ow and heat transfer characteris-
tics due to lid driven mixed convection within porous square cav- Authors would like to thank anonymous reviewers for critical
ities has been studied in the present investigation. The penalty comments and suggestions which improved the quality of the
nite element method helps to obtain smooth solutions in terms manuscript.
5448 D. Ramakrishna et al. / International Journal of Heat and Mass Transfer 55 (2012) 54365448

References uid-saturated porous medium, Numer. Heat Transfer Part A-Applications

53 (2008) 88110.
[20] S. Kimura, A. Bejan, The heatline visualization of convective heat-transfer, J.
[1] V.E. Schrock, A.D.K. Laird, Physical modeling of combined forced and natural
Heat Transfer Trans. ASME 105 (1983) 916919.
convection in wet geothermal formations, J. Heat Transfer Trans. ASME 98
[21] A. Bejan, Convection Heat Transfer, Wiley, New York, 1984.
(1976) 213220.
[22] F.Y. Zhao, D. Liu, G.F. Tang, Natural convection in a porous enclosure with a
[2] Y. Boutin, L. Gosselin, Optimal mixed convection for maximal energy recovery
partial heating and salting element, Int. J. Therm. Sci. 47 (2008) 569583.
with vertical porous channel (solar wall), Renew. Energy 34 (2009) 2714
[23] Y. Varol, H.F. Oztop, M. Mobedi, I. Pop, Visualization of natural convection heat
2721.
transport using heatline method in porous non-isothermally heated triangular
[3] W.L. Pu, P. Cheng, T.S. Zhao, An experimental study of mixed convection heat
cavity, Int. J. Heat Mass Transfer 51 (2008) 50405051.
transfer in vertical packed channels, AIAA J. Thermophys. Heat Transfer 13
[24] E. Hakyemez, M. Mobedi, H.F. Oztop, Effects of wall-located heat barrier on
(1999) 517521.
conjugate conduction/natural-convection heat transfer and uid ow in
[4] R. Rachedi, S. Chikh, Enhancement of electronic cooling by insertion of foam
enclosures, Numer. Heat Transfer Part A-Applications 54 (2008) 197220.
materials, Heat Mass Transfer 37 (2001) 371378.
[25] Q.H. Deng, Fluid ow and heat transfer characteristics of natural convection in
[5] I. Muhaimin, R. Kandasamy, A.B. Khamis, Numerical investigation of variable
square cavities due to discrete source-sink pairs, Int. J. Heat Mass Transfer 51
viscosities and thermal stratication effects on MHD mixed convective heat
(2008) 59495957.
and mass transfer past a porous wedge in the presence of a chemical reaction,
[26] A. Dalal, M.K. Das, Heatline method for the visualization of natural convection
Appl. Math. Mech. Engl. Ed. 30 (2009) 13531364.
in a complicated cavity, Int. J. Heat Mass Transfer 51 (2008) 263272.
[6] O.G. Martynenko, P.P. Khramtsov, Free-Convective Heat Transfer, Springer,
[27] R.S. Kaluri, T. Basak, Heatline analysis of thermal mixing due to natural
Berlin, 2005.
convection in discretely heated porous cavities lled with various uids,
[7] D.A. Nield, A. Bejan, Convection in Porous Media, second ed., Springer, New
Chem. Eng. Sci. 65 (2010) 21322152.
York, 1999.
[28] J.N. Reddy, An Introduction to Finite Element Analysis, McGraw-Hill, New
[8] D.B. Ingham, I. Pop (Eds.), Transport Phenomena in Porous Median, Pergamon,
York, 1993.
Oxford, 1998.
[29] T. Basak, P.V.K. Pradeep, S. Roy, I. Pop, Finite element based heatline approach
[9] D.B. Ingham, I. Pop (Eds.), Transport Phenomena in Porous Median, Pergamon,
to study mixed convection in a porous square cavity with various wall thermal
Oxford, 2002.
boundary conditions, Int. J. Heat Mass Transfer 54 (2011) 17061727.
[10] D.B. Ingham, A. Bejan, E. Mamut, I. Pop (Eds.), Emerging Technologies and
[30] T. Basak, S. Roy, T. Paul, I. Pop, Natural convection in a square cavity lled with
Techniques in Porous Media, Kluwer, Dordrecht, 2004.
a porous medium: Effects of various thermal boundary conditions, Int. J. Heat
[11] K. Vafai (Ed.), Handbook of Porous Media, Marcel Dekker, New York, 2000.
Mass Transfer 49 (2006) 14301441.
[12] K. Vafai (Ed.), Handbook of Porous Media, second ed., Taylor and Francis, New
[31] T. Basak, S. Roy, Role of Bejans heatlines in heat ow visualization and optimal
York, 2005.
thermal mixing for differentially heated square enclosures, Int. J. Heat Mass
[13] I. Pop, D.B. Ingham, Convective Heat Transfer: Mathematical and
Transfer 51 (2008) 34863503.
Computational Modelling of Viscous Fluids and Porous Media, Pergamon,
[32] T. Basak, S. Roy, I. Pop, Heat ow analysis for natural convection within
Oxford, 2001.
trapezoidal enclosures based on heatline concept, Int. J. Heat Mass Transfer 52
[14] A. Bejan, I. Dincer, S. Lorente, A.F. Miguel, A.H. Reis, Porous and Complex Flow
(2009) 24712483.
Structures in Modern Technologies, Springer, New York, 2004.
[33] H.F. Oztop, I. Dagtekin, Mixed convection in two-sided lid-driven differentially
[15] K.M. Khanafer, A.J. Chamkha, Mixed convection ow in a lid-driven enclosure
heated square cavity, Int. J. Heat Mass Transfer 47 (2004) 17611769.
lled with a uid-saturated porous medium, Int. J. Heat Mass Transfer 42
[34] M. Sathiyamoorthy, T. Basak, S. Roy, I. Pop, Steady natural convection ow in a
(1999) 24652481.
square cavity lled with a porous medium for linearly heated side wall(s), Int.
[16] H.F. Oztop, Combined convection heat transfer in a porous lid-driven enclosure
J. Heat Mass Transfer 50 (2007) 18921901.
due to heater with nite length, Int. Commun. Heat Mass Transfer 33 (2006)
[35] G. Lauriat, V. Prasad, Non-Darcian effects on natural convection in a vertical
772779.
porous enclosure, Int. J. Heat Mass Transfer 32 (1989) 21352148.
[17] K. Khanafer, K. Vafai, Double-diffusive mixed convection in a lid-driven
[36] M.K. Moallemi, K.S. Jang, Prandtl number effects on laminar mixed convection
enclosure lled with a uid-saturated porous medium, Numer. Heat Transfer
heat transfer in a lid-driven cavity, Int. J. Heat Mass Transfer 35 (1992) 1881
(Part A) 42 (2002) 465486.
1892.
[18] P. Kandaswamy, M. Muthtamilselvan, J. Lee, Prandtl number effects on mixed
[37] Q.H. Deng, G.F. Tang, Numerical visualization of mass and heat transport for
convection in a lid-driven porous cavity, J. Porous Media 11 (2008) 791801.
mixed convective heat transfer by streamline and heatline, Int. J. Heat Mass
[19] E. Vishnuvardhanarao, M.K. Das, Laminar mixed convection in a parallel
Transfer 45 (2002) 23872396.
two-sided lid-driven differentially heated square cavity lled with a