International Journal of Heat and Mass Transfer 111 (2017) 329–345

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

MHD convection and entropy generation of nanofluid in a porous

enclosure with sinusoidal heating
Sumit Malik, A.K. Nayak ⇑
Department of Mathematics, Indian Institute Technology of Roorkee, Roorkee 247667, India

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

Article history: A numerical study of heat transfer and entropy generation of a magneto-hydrodynamic (MHD) nanofluid
Received 7 November 2016 flow inside an enclosure filled with a fluid saturated porous medium is presented. The flow is influenced
Received in revised form 6 March 2017 by time periodic discrete heat sources along the short side walls. A detailed physical insights of time
Accepted 31 March 2017
dependent flow and heat transfer is presented based on various flow governing parameters such as
Grashof number (104 —106 ), Hartmann number (1–50), Darcy number (0.001–1.0) and nanoparticle vol-
ume fraction (0.0–0.20) with a fixed Prandtl number (6.2). The resulting energy flux vectors are simulated
Keywords:
to analyze the convection generated heat transfer ratio. Entropy generation and Bejan number are used to
Cu-water nanofluid
Porous mixture
study the performance of the system.
Magnetic convection Ó 2017 Elsevier Ltd. All rights reserved.
Energy flux vector
Entropy generation
Bejan number

1. Introduction The effect of MHD convection in an electrically conducting fluid

Large scale applications such as magneto-hydrodynamic gener-
ators, electrical equipment cooling, plasma physics, geothermal
energy extraction, the chemical pollutants spreading in saturate
soil and boundary layer control theory in the field of mechanics
requires a better understanding of Magneto-hydrodynamic flow
of an electrically-conducting fluid coupled with heat transfer [1].
Many engineering applications deal with the heat transfer by nat-
ural convection in enclosures such as cooling of electronic devices,
solar energy, heat exchanger design, continuous strips or filaments
[2,3]. In practical applications, the process of heating and cooling of
fluid is very sensitive and sometimes requires very quick response
of heating as well as cooling. Clear fluid like water and ethylene
glycol fails this quick response property to heating and cooling
due to which the nanoparticles with relatively high thermal con-
ductivity such as Cu, Al2O3, and TiO2 are required to be suspended
into the base fluid. Now a days, the MHD flow inside a porous med-
ium has drawn the interest of several researchers because of its
importance in many industrial applications such as MHD power
generator designing, control of chemical waste and pollutants dis-
semination, optimization of solidification processes of metals and
alloys. The relative problem also appear in the operation of micro-
electronic devices and electronic packages [4].
⇑ Corresponding author.
E-mail address: ameeyakumar@gmail.com (A.K. Nayak).
inside a differentially heated rectangular enclosure is investigated
by Rudraiah et al. [5]. They found that the convective heat transfer
decreases with increasing Hartmann number. Oztop et al. [6] stud-
ied a square cavity with two semicircular heaters placed at the
lower wall considering adiabatic horizontal walls and isothermal
boundary conditions on vertical walls in presence of magnetic
field. They concluded that the flow strength as well as convective
heat transfer decreases with increasing magnetic field effect. Siva-
sankaran et al. [7] used finite volume method to study the effect of
sinusoidal boundary temperatures at the vertical walls in a lid dri-
ven square cavity having adiabatic horizontal walls. They con-
cluded that the amplitude ratio increases the heat transfer rate.
The work is further extended in the presence of magnetic field
and found that increment in Hartmann number decreases the total
heat transfer rate in Buoyancy-driven convection [8]. Subse-
quently, Sivasankaran and Pan [9] used saturated porous medium
with same boundary conditions to investigate the effect of ampli-
tude ratio on the heat transfer rate. The temperature difference is
found to be increasing with increase of amplitude ratio, Darcy
number and porosity.
Mostly, nanofluids are used to obtain higher heat transfer rate
in industry compared to low conductivity clear fluid. The term
‘nanofluid’ was coined by Choi and Eastman [10] and a number
of experimental as well as numerical studies have been conducted
by considering the thermal conductivity of nanofluids till date.
Experimentally, Eastman et al. [11] used 0.3% volume fraction of

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2017.03.123
0017-9310/Ó 2017 Elsevier Ltd. All rights reserved.
330 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345
Nomenclature
Cp specific heat capacity (J/kg K)
g gravitational acceleration (m=s2 )
k thermal conductivity (W/m K)
x
; y
Cartesian coordinates (m)
x; y dimensionless Cartesian coordinates
u
; v
components of velocity in x
and y
directions (m/s)
u; v dimensionless of velocity components in x and y
directions
t
time (s)
t dimensionless time
p
pressure (N m2 )
p dimensionless pressure
B0 magnetic field (Amp/m)
K permeability of porous medium (m2 )
T
temperature (K)
T dimensionless temperature
L cavity length (m)
S
entropy generation
S dimensionless entropy generation
Ra Rayleigh Number
Pr Prandtl number
Gr Grashof Number
Nu Nusselt number
H0 heat function
E energy flux vector
copper nanoparticles of size <10 nm in ethylene glycol and the
effective thermal conductivity is found to be increased up to 40%
as compared to pure ethylene glycol. Recently, Ghodsinezhad
et al. [12] performed an experimental study to find the optimum
fraction of nanoparticles in natural convection of Al2O3-water
nanofluids. In natural convection, it is predicted that heat transfer
rate is increasing up to 15% if the volume fraction is changed from
0% to 0.1% and further increment of nanoparticle fraction descends
the rate of heat transfer. Recently, numerical study of mixed con-
vection of Cu-water nanofluid in a two-sided lid-driven square
enclosure is conducted by Malik and Nayak [13] using two discrete
heat sources along the bottom wall. They concluded that the heat
transfer rate is sensitive to nanoparticle volume fraction and it
increases with increase in nanoparticle volume fraction. The effect
of magnetic field on heat and fluid flow in a CuO-water nanofluid
filled enclosure heated from below is studied by Sheikholeslami
et al. [14]. They found that heat transfer is increasing with the
increment of heat source length and Hartmann number, and
decreases with increment in Rayleigh number. Ghasemi et al.
[15] studied the natural convection of Al2O3-water nanofluid inside
a differentially heated square enclosure in presence of magnetic
field. The results show that increment of nanoparticle volume frac-
tion results in the variation of heat transfer depending on the
increment of Hartmann number and increment of Rayleigh num-
ber respectively.
MHD flow in nanofluid filled domain is an important aspect of
study as it is beneficial for energy optimization and designing of
heat removal devices of MHD power generator [16,17]. In a recent
study, Elshehabey and Ahmed [18] presented a numerical study of
MHD mixed convection of nanofluid inside a square enclosure with
sinusoidal temperature and nanoparticle volume fraction on both
vertical walls. The results show that the flow variation inside the
cavity decreases with increasing the buoyancy ratio. Pekmen and
Tezer-Sezgin [19] numerically studied the mixed convection flow
inside a lid driven square enclosure filled with porous medium
activated by magnetic field using local thermal equilibrium(LTE)
Greek letters
q density (kg=m3 )
l dynamic viscosity (Pa s)
r electrical conductivity (lS=cm)
b coefficient of volume expansion (K1 )
a thermal diffusivity (k=ðqC p Þ) (m2 =s)
/ nanoparticle volume fraction
s dimensionless time
w stream function
Subscripts
nf nanofluid
f fluid
H hot
C cold
s solid
ht due to heat transfer
ff due to fluid flow
mf due to magnetic field
m average
o reference state
approach. They have concluded that the fluid flow becomes slower
with the increase of the magnetic field intensity and decrease in
porosity. In LTE approach the fluid temperature considered to be
equal to the solid temperature, whereas in local thermal non-
equilibrium (LTNE) approach different temperature of fluid and
solid medium is considered. LTE approach is used by many authors
for transport phenomena in porous media such as Al-Nimr and
Haddad [20], Rees and Vafai [21], Kim and Vafai [22] and Mohamad
[23]. Haddad et al. [24] made an analytical treatment for the vali-
dation of LTE assumption of a natural convective flow in a porous
media along a vertical flat plate. A lot more high quality research
are also performed on the local thermal non-equilibrium (LTNE)
approach in nanofluids which is also an advanced model for the
flow of nanofluid in porous medium [25–29].
The analysis of entropy generation in magnetohydrodynamic
porous channel is made by Ibáñez [30] analytically. Their result
shows that the total entropy generated during the process
increases with slip flow due to decrease in the temperature of
the system. Khalili et al. [31] investigated the unsteady MHD flow
and heat transfer over a stretching/shrinking sheet in a nanofluid
filled porous medium and reported that increase in nanoparticle
volume fraction results in increase of skin friction and heat transfer
coefficient. Recently, Kefayati [32] discussed the effect of non-
Newtonian nanofluids on heat transfer and entropy generation in
a porous cavity and concluded that increment in Rayleigh number
increases the heat transfer and entropy generation. In another
recent study Kefayati and Sidik [33] studied the non-Newtonian
nanofluid in an inclined cavity for the effect of entropy generation.
The result shows enhancement in fluid friction and heat transfer
irreversibility with increasing Rayleigh number. Fersadou et al.
[34] made a numerical study to analyze the effect of magnetic field
on entropy generation inside a vertical porous channel filled with
Cu-water nanofluid. Results show that nanoparticle addition
increases the entropy generation of the system. The entropy gener-
ation in a partially heated two sided lid driven cavity under mag-
netic field effect filled with Al2O3-water nanofluid is studied by
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345 331

Hussain et al. [35]. The augmentation of total entropy generation

due to increase in Reynolds number and solid volume fraction is
analyzed.
In the present study, the convective heat transfer is described
by using energy flux vectors. The conductive heat transfer is easily
visualized using the isotherms whereas convective mode of heat
transfer can be understood by using of heatlines. It is also reported
that visualization of heat transfer rate due to heatlines compare to
isotherms are better in case of convective mode of transfer rate
[36]. Kaluri and Basak [37] presented the heatline analysis of nat-
ural convection due to discrete heating in porous cavities for a
wide range of Prandtl number, Rayleigh number and Darcy num-
ber. The study of heatlines is found to be very useful in the analysis
of heat transfer in discretely heated enclosures filled with porous
medium. For smaller Darcy number a conduction dominated heat
transfer is reported whereas heat transfer is found to be convection
dominated for higher Darcy number. Heatline analysis of convec-
tive heat transfer in differentially heated square enclosure filled
Fig. 1. Physical configuration. T H represents the hot region and T C represents cold
with nanofluid is made by Basak and Chamkha [38]. The heat region of the wall.
transfer due to convection is found to be dominant in nanofluids.
Recently, Hooman [39] suggested that instead of solving a difficult
form of heat function equations, it is easy to use the concept of Cu-water nanofluid. Constant temperature has been maintained
energy flux vectors, which are locally tangent to the heatlines. at the mid portion of right vertical walls of length L=3 and its right
The heat transfer lines are not sufficient enough to visualize energy counterpart of the same length is maintained with a time-periodic
transfer effect in case of a transient flow and hence in the present heat transfer of oscillating nature. The remaining sections of the
study energy flux vectors are used to visualize the flow of energy. boundaries are assumed to be thermally insulated. The fluid is con-
Moreover, the energy flux vectors can also serve the purpose of sidered to be Newtonian fluid which is viscous and incompressible
reducing extra efforts of formulating and solving the additional dif- and a laminar flow which is in thermal equilibrium with negligible
ferential equations without implementing the boundary conditions slip effect between two phases has been assumed. Boussinesq
as in heat lines [40]. assumption is followed for variation of density. The shape of
Nithyadevi et al. [41] considered the natural convective flow nanoparticles is assumed to be spherical with 100 nm diameter.
inside a square enclosure with partially active thermal zones The magnetic Reynolds number (Rem ¼ lf rf um L  1) is assumed
applied along the vertical walls in which one is time periodic and to be small which enables the induced magnetic effect to be trivial
the other is isothermal. The rate of heat transfer is found to be as compared to the effect of external magnetic field [43]. The ther-
highest for the middle-middle thermally activated zones. The work mophysical properties [44] of nanofluid are tabulated in Table 1.
is further extended by Nayak et al. [42] by considering the enclo- The flow in the porous medium is dealt with Darcy-Brinkman-
sure filled with porous medium in the presence of magnetic field. Forchheimer model. The saturated porous medium is assumed to
The study shows higher heat transfer rate in middle-middle com- be isotropic thermally, electrically and hydrodynamically. The Hall
bination of short side heated and cold regions as compared to effect, viscous dissipation, radiation effect, Joule heating and pres-
the other combinations and the increase in Prandtl number results sure work are also assumed to be negligible [42,45].
in increase in the heat transfer rate for Da < 1. In the present work,
the author tried to find the optimum heat transfer and entropy 2.2. The governing equations
generation for natural convection of nanofluid flow in enclosure
containing saturated porous media in the presence of magnetic Equations governing the two dimensional flow that satisfies the
field. To the best of our knowledge, the problem dealing with nano- above mentioned assumptions are given by [46,47],
fluid flow inside a saturated porous medium in the presence of Continuity equation
magnetic field with periodic heat transfer along the short side
@u
@ v

walls has not been studied so far. The effect of various flow govern- þ ¼ 0; ð1Þ
@x
@y

ing parameters such as Grashof number, Hartmann number, Darcy
number and nanoparticle volume fraction over the fluid flow is Momentum equations
studied. The study focuses on the optimum heat transfer and " !
@u

@u



@u


1 @p
@ 2 u
@ 2 u

entropy generation analysis by variations of above mentioned þu þv ¼ 
þ lnf þ
parameters. The fluid flow and heat transfer behavior inside the @t
@x
@y
qnf @x @x
2 @y
2
enclosure and convective heat transfer analysis is made by the help lnf u

of energy flux vectors. The time history analysis is investigated to rnf B20 u
 ; ð2Þ
K
visualize the flow enhancement during the initial time level and
the flow phase variation with respect to time.

Table 1
Thermophysical properties of water and copper [44].
2. Problem formulation
Property Water Copper
2.1. Flow configuration Cp 4179 383
q 997.1 8954
The physical domain as represented in Fig. 1 considers a k 0.6 400
b 2:1  104 1:67  105
two-dimensional closed square enclosure of length L, filled with
r 0.05 5:96  107
a saturated porous medium containing an electrically conducting
332 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345

" !
@v

@v



@v


1 @p
@2v
@2v
and

þu þ v ¼  þ l þ
@t @x
@y
qnf @y
nf
@x
2 @y
2 rnf B20
 S
mf ¼ u
2 ð16Þ
lnf v
T
0
þðqbÞnf gðT
 T
C Þ  ; ð3Þ
K T
þT

with, T
0 ¼ H 2 C .
and energy equation Using the non-dimensional variables [46,47,54]
!
@T

@T



@T


@2T
@2T
u
L v
L x
y
p
L2
þu þv ¼ anf þ : ð4Þ u¼ ; v¼ ; x¼ ; y¼ ; p¼ ;
@t
@x
@y
@x
2 @y
2 af af L L qnf a2f
The associated boundary conditions are, t
af T
 T
C
t¼ ; T¼ kf ;
L 2 ðT
H  T
C ÞL
@T

u
¼ v
¼ ¼ 0 at y
¼ 0 and L; 0 6 x
6 L
@y

gbf ðT H  T C ÞL3 q2f lf
Gr ¼ ; Pr ¼ ;
u
¼ v
¼ 0 at x
¼ 0 and L; l qf af
2
0 6 y
6 L f
sffiffiffiffiffi
with rf K
Ha ¼ B0 L and Da ¼ 2 ; ð17Þ
 lf L
T
H ð1

 asinð2pt =xÞÞ at x ¼ 0;

T
¼ L=3 6 y
6 2L=3
T
C at x
¼ L the dimensionless form of the governing equations can be written
as,
and
@u @ v
@T
þ ¼ 0; ð18Þ
¼ 0 elasewhere on x
¼ 0 and x
¼ L ð5Þ @x @y
@x

!
where the effective density and dynamic viscosity are defined by @u @u @u @p lnf @2u @2u
[48–50], þu þv ¼ þ þ
@t @x @y @x qnf af @x2 @y2
qnf ¼ /qs þ ð1  /Þqf ; ð6Þ rnf bf lnf u
 uHa2 Pr  ; ð19Þ
lf rf qnf qnf af Da
lnf ¼ ; ð7Þ
!
ð1  /Þ2:5
@v @v @v @p lnf @2v @2v
þu þv ¼ þ þ
and the thermal diffusivity, heat capacitance and thermal expansion @t @x @y @y qnf af @x2 @y2
coefficient are given by,
ðqbÞnf lnf v
k þ GrPr 2 T  ; ð20Þ
anf ¼ nf ; ð8Þ qnf bf qnf af Da
ðqC p Þnf
!
@T @T @T anf @2T @2T
ðqC p Þnf ¼ /ðqC p Þs þ ð1  /ÞðqC p Þf ; ð9Þ þu þv ¼ þ ; ð21Þ
@t @x @y af @x2 @y2
ðqbÞnf ¼ /ðqbÞs þ ð1  /ÞðqbÞf : ð10Þ The dimensionless form of boundary conditions are given by,
The effective thermal conductivity and electrical conductivity for @T
u¼v¼ ¼ 0 at y ¼ 0 and 1; 0 6 x 6 1;
uniformly distributed spherical shaped nanoparticles are calculated @y
using Maxwell model [51] as,
u ¼ v ¼ 0 at x ¼ 0 and 1; 06y61
knf ks þ 2kf  2/ðkf  ks Þ
¼ ð11Þ
kf ks þ 2kf þ /ðkf  ks Þ with

1  Asinð2pt=dÞ at x ¼ 0;
rnf 3ðrrs  1Þ/ T¼ 1=3 6 y 6 2=3;
¼ 1 þ rs f
ð12Þ 0 at x ¼ 1
rf ðr þ 2Þ  ðrrs  1Þ/
f f
and
The performance of the system is reduced by various factors that
@T
creates irreversibility in the system. There are three sources of ¼ 0 elasewhere on x ¼ 0 and x ¼ 1 ð22Þ
@x
entropy generation in our system: the heat transfer, viscous effects
and the magnetic field. Hence, the total entropy generation rate is where A and 2d denotes the amplitude factor and the period
given by [47,52,53], respectively.
For time t ¼ 0, the initial conditions are taken as follows,
S
gen ¼ S
ht þ S
ff þ S
mf ð13Þ
u ¼ v ¼ T ¼ 0 for 0 6 x 6 1; 0 6 y 6 1: ð23Þ
where
" 2  #

2 Initially, the physical variables are assumed to be
knf @T
@T
S
ht ¼ þ ð14Þ
T
2 @x
@y
Sgen ¼ Sht þ Sff þ Smf ; ð24Þ
0

" (  where
2 
2 
2 )#
lnf K qnf @u
@v @u @v
" 2  2 #
S
ff ¼ ðu þ v Þ þ

2
2
2 þ2 þ þ knf @T @T
T
0 lnf @x
@y
@y
@x
Sht ¼ þ ; ð25Þ
kf @x @y
ð15Þ
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345 333

" (    2  2 )# Z Z Z Z Z Z
@v @u @ v @u @u @u
2 n e n e e n
lnf @u v dydx
Sff ¼ v ðu þ v Þ þ Da 2
2 2
þ2 þ þ ; dxdy þ u dxdy þ
lf @x @y @y @x s w @t s w @x w s @y
Z nZ e
@p
ð26Þ ¼ dxdy
s w @x
"Z Z Z eZ n 2 #
rnf Ha2 2 lnf n e 2
@ u @ u
Smf ¼ v u ð27Þ þ dxdy þ dydx
rf qnf af s w @x2 w s @y2
" #Z Z
and rnf bf lnf n e
 Ha2 Pr þ udxdy: ð36Þ
! rf qnf qnf af Da s w
2
lf T 0 af
v¼ ; ð28Þ The reduced form of the equation can be represented as,
kf KðT H  T C Þ2
F e ue  F w uw þ F n un  F s us ¼ b ð37Þ
where v is called as irreversibility distribution ratio. The value of v
where F 0 s denote the non-linear coefficient of u0 s and the source and
is considered to be 0.01 in all the computations [55–57]. The vol-
time derivative terms are contained in b.
ume averaged entropy generation is calculated as,
A QUICK (quadratic upstream interpolation for convective kine-
Z matics) third-order accurate scheme is used for the discretization
1
Sm ¼ Sgen dV; ð29Þ of convective terms of the governing equations. For instance, at
V V interface the approximation due to quadratic interpolation for u is,

where V is the total nanofluid volume. 1

ue ¼ ð3uE þ 6uP  uW Þ if F e > 0; ð38Þ
The relative dominance of heat transfer irreversibility is ana- 8
lyzed by using Bejan number which is calculated as the ratio of
heat transfer entropy generation to the total entropy generation 1
ue ¼ ð6uE þ 3uP  uW Þ if F e < 0; ð39Þ
and is written as, 8
and hence,
Sht
Be ¼ : ð30Þ 1
Sgen F e ue ¼ ð3uE þ 6uP  uW Þ½½F e ; 0
8
The heat transfer rate is characterized by local (Nu) and average 1
 ð6uE þ 3uP  uW Þ½½F e ; 0 ð40Þ
Nusselt number (Num ) and are represented as, 8
Similar procedure is followed for the computations at west
 
knf @T north and south faces and time derivatives are discretized using
Nuðx ¼ 0Þ ¼  ; ð31Þ
kf @y a third-order implicit scheme. The system of discretized pressure
linked equations are solved iteratively by using SIMPLE (Semi
R Implicit Method for Pressure Linked Equations).
Nuðx ¼ 0Þdy
Num ðx ¼ 0Þ ¼ R : ð32Þ The resultant discretized system of equations is dealt with
dy using the iterative Semi Implicit Method for Pressure Linked Equa-
The heat flow inside the enclosure is visualized due to energy flux tion (SIMPLE algorithm) based on the pressure correction
vectors are represented by, approach. SIMPLE method inherits the idea of finite volume meth-
ods. The procedure of finding the solutions of successive steady
@H0 ^ @H0 ^ state problems is the initial step that gradually pave the way for
E¼ i j; ð33Þ
@y @x the solution of current unsteady problem. For any reasonable value
  of Dt, the values of the unknown dependent variable U at time
where H0 is calculated from conductive heat fluxes  @T
@x
;  @T
@y
and t þ Dt can be approximated using the values of U at time t taken
convective heat fluxes (uT; v T), which satisfies the steady energy as initial guess. This iterative procedure is carried out until a con-
balance equation, vergent solution is obtained. The flow field variables are required
to move through a progression of smaller time-steps Dt ¼ 0:001
@H0 anf @T @H0 anf @T to ultimately obtain a repeated time independent and convergent
¼ uT  ; ¼ v T þ : ð34Þ solution. The steady state condition considering t ¼ 10 has been
@y af @x @x af @y
adopted to compute all the numerical results which have been pre-
sented in the form of figures, as the values are repeated. The fol-
lowing sequence has been followed to obtain the solution of a
3. Numerical approach
particular iteration that consists of pressure dependent terms:
3.1. Methodology
1. The discretized form of flow governing and heat transfer equa-
tions provides an algebraic system of equations that can be
All the computations are performed using finite volume method
written as follows:
based on SIMPLE algorithm [58]. The system of non-dimensional
governing Eqs. (18)–(21) is integrated over the stated control vol- AU Unþ1 nþ1
i1;j þ BU Ui;j þ C U Unþ1
iþ1;j ¼ DU ð41Þ
ume as,
Z Z where U is used for u; v and h and AU , BU , and C U together forms
@u @v
dV þ dV ¼ 0; ðcontinuityÞ ð35Þ the coefficient matrix and DU represents the pressure and source
DV @x DV @y
terms for U at nth time step. This system of equations in turn can
where DV denotes the volume of the nodal point. The momentum be written in the form of matrices in which the coefficient
equation can be represented as, matrix is in tri-diagonal form.
334 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345
2. As the energy and momentum equations can be coupled, a
block elimination procedure has been employed to deal with
this system of equations.
3. The Successive under relaxation method is employed to solve
associated pressure equation that is an analog of Poisson
equation.
4. The pressure correction is used to update the velocity field at
each cell.
The following convergence criteria is employed throughout the
computations,
i;j  ei;j j < 
jenþ1
n
where e is used for u, v or T, n is the time level, i; j represent the cell
index and  is the tolerance which is taken as 105 .
3.2. Grid independency and code validation
The code validation and grid independency test is presented in
Fig. 2(a) by considering the grid size between 81  81 and
121  121 and compared with the results of Ghasemi et al. [15].
The results show the variation of average Nusselt number with
varying Hartmann number from 0 to 60 for Ra ¼ 105 , / ¼ 0:03
and Pr ¼ 6:2 in natural convection flow of nanofluids in the pres-
ence of magnetic field. The results did not change for grid size of
101  101 and higher. Hence, for the present study the grid size
of 101  101 is chosen for all computations. Another code valida-
tion test is conducted and compared with the outcomes of
Sathiyamoorthy et al. [59] for average Nusselt number with vary-
ing Rayleigh number from 103 to 106 for Da ¼ 103 and Pr ¼ 0:7
and represented in Fig. 2(b). The results show the variation of aver-
age Nusselt number with Rayleigh number for clear fluid in porous
domain. Our results found to be in excellent agreement with those
of Sathiyamoorthy et al. Ho et al. [60] made an experimental study
on Al2 O3 -water nanofluid, the results are presented in the form of
average Nusselt number at varying Rayleigh numbers for different
nanoparticle volume fractions. Fig. 2(c) shows the comparison of
average Nusselt number at different Rayleigh numbers with
/ ¼ 0:001. The results are found to be in good agreement with
the experimental results by Ho et al. [60]. The comparison for
the validation of entropy generation is made with the results of
Basak et al. [57]. The study of entropy generation in a porous cavity
is made by considering five different cases of natural convection in
square cavity. The presented comparison in Fig. 2(d) is made with
case 5 of the study and the results are found to be in a good agree-
ment with [57].
4. Results and discussion
The MHD study of electrically conducting nanofluid flow in a
porous medium is presented with discrete time periodic heat
source along the left vertical wall whereas the opposite wall (right)
is maintained at a constant but lower temperature. The results are
presented with the variation of Grashof number (104 6 Gr 6 106 ,
to measure the thermal buoyancy effects in natural convection),
Hartmann number (1 6 Ha 6 50, to measure the magnetic field
effects), Darcy number (0:001 6 Da 6 1:0, measuring the effect of
porosity) and / (0:0 6 / 6 0:2, to measure the effect of nanoparti-
cle volume fraction on heat transfer) to obtain the effective heat
transfer with a fixed Prandtl number (Pr ¼ 6:2). The outcomes
are discussed using streamlines, isotherms, energy flux vectors,
Num (to measure the average heat transfer near the heat source),
Sm (to measure the total entropy release) and Be (measuring the
dominance of irreversibility due to heat transfer over the irre-
versibility due to fluid friction and magnetic field).
The transient behavior of fluid flow and heat transfer enhance-
ment due to natural convection at various time levels is presented
in the form of streamlines and isotherms in Fig. 3 for Gr ¼ 104 ,
Ha ¼ 10, Da ¼ 0:001 and / ¼ 0:0. As the process starts the fluid
adjacent to the time periodic heated zone on left vertical wall gets
activated. A small clockwise rotating circulation zone is formed
closed to the heat source due to conduction effect. With the
increase of time the circulation zone move towards the right sec-
tion of the enclosure and stacked at the cold part of the enclosure.
It is observed that, as time proceeds the heat transfer increases
ð42Þ
resulting the increment of flow velocity. At t ¼ 0:05, the stream
vortex formed a vertical oval shape adjacent to the hot wall and
moves towards the right section as time increases. The vortex
becomes stronger as time increases to t ¼ 0:30 representing a cir-
cular shape along the central part of the enclosure defines a shift of
heat transfer mode from conduction to convection. For higher time
levels the shape deformed to a elliptical shape is representing a
fully convection dominated mode of heat transfer.
The isotherms at varying time spans represents the correspond-
ing growth of rate of heat transfer with increase in time. At small
time levels, it is observed that a semi-elliptical thermal zone is
formed close to the thermally activated region due to the effect
of thermal boundary layer. From Fig. 3, the movement of heat
transfer lines value at 0.20 is clearly observed as time increases
from t ¼ 0:10 to t ¼ 0:15. For further increment of time the iso-
therms expand and move towards the central zone and the ther-
mal boundary layer has less influence on heat transfer shows a
convection mode of heat transfer rate. A secondary vortex starts
to appear along with the right cold wall at t ¼ 0:25 and the heat
transfer of fluid is getting stronger and achieved a steady state
when the isotherms are coincides with each other. No further
improvement of heat transfer is observed around the right vertical
wall after t ¼ 0:50, i.e., always represents the same value closed to
the right wall.
Fig. 4 represents the streamlines, isotherms and energy flux
vectors for Gr ¼ 104 , 105 and 106 at Ha ¼ 50, Da ¼ 0:001 and
/ ¼ 0:2 (dashed lines) when length of heat source is L=3. Solid lines
in the streamlines and isotherm curves corresponds to / ¼ 0:0
(clear fluid). As the temperature of the mid active location is larger
as compared to the other sections of the wall, the fluid close to the
hot region gets heated and loose its density. Weak dense fluid rises
above its mean level and forced towards the cold locations of the
right vertical wall resulting a faster movement of fluid particles.
This movement from hot region to cold region of fluid particles
forms a large recirculating clockwise rotating eddy inside the
enclosure within which a smaller clockwise eddy is formed with
a small stagnant region to avoid the shear instability as repre-
sented in Fig. 4(a) and (d). In Fig. 4(a) the streamlines correspond-
ing to / ¼ 0:2 appears inside the streamlines corresponding to
/ ¼ 0:0, i.e., most of the clear fluid occupies outer region of the cav-
ity but with the increase in nanoparticle volume fraction, the shear
force increases due to which the fluid occupies the core of enclo-
sure. The isotherms lines corresponding to / ¼ 0:0 are limited to
low values of streamlines as compared to that of / ¼ 0:2, which
shows higher heat transfer rate in case of nanofluids as compared
to clear fluid. As Gr increases to 105 the streamlines corresponding
to clear fluid in the core of the enclosure forms an oval shape, and
appears closer to heated region as compared to the streamlines of
nanofluid, which is due to the higher effect of buoyancy forces on
nanofluid. With further increase of Gr values, it is found that two
small eddies rotating clockwise are formed with a small stagnant
fluid (Fig. 4(g)). The stagnant fluid region started to form a circular
shape for Gr ¼ 104 , elongated to elliptical shape for Gr ¼ 105 and
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345 335

5
Ghasemi et al.[2011]
∗ 1.5

4.5 121 x 121 Sathiyamoorthy et al.[2007]

∗ 101 x 101 Present Result
4 ∗ 81 x 81

1
3.5
Num

Num
3 ∗

0.5
2.5
∗
2
∗
1.5 0 3 4 5 6
0 10 20 30 40 50 60 10 10 10 10
Ha Ra
(a) (b)

11 30

Ho et al. [2010] Basak et al.[2012]

25
10 Present Result Present Result

20
9
Num

Sm

15

8
10

7
5

6 0 -5
1E+06 2E+06 3E+06 10 10-4 10-3
Ra Da
(c) (d)
Fig. 2. Comparison of average Nusselt number with (a) the results of [15] for Ra ¼ 105 , / ¼ 0:03, Pr ¼ 6:2 and 0 6 Ha 6 60 for various grid size from 81  81 to 121  121, (b)
the results of [59] for Da ¼ 103 , Pr ¼ 0:7 and 103 6 Ra 6 106 , (c) the results of [60] for 6  105 6 Ra 6 3:37  106 at / ¼ 0:001, and (d) the results of [57] for Ra ¼ 5  105 with
105 6 Da 6 103 at Pr ¼ 10.

splits into two for Gr ¼ 106 representing the buoyancy dominated
effect as compared to a natural convection effects.
Fig. 4(b), (e), and (h) represent the maximum variation of heat
transfer lines for the case of Gr ¼ 106 , the isotherms at mid section
becomes almost parallel to the horizontal walls of the enclosure
showing the convection dominated flow in the corresponding
region. When the heater on the left wall is switched on, left wall
attains the maximum temperature of the mid-plane. A major con-
vection effect is indicated by the energy flux vectors towards the
right wall from left part of the enclosure (4(c), (f), and (i)). This is
different from steady state outcomes where most of the energy
flux vectors are released from the left heating portion. With peak
4 6
shifts of Grashof number from 10 to 10 , the output pattern of
heat transfer lines increasingly changes close to the heater to over-
come the sinusoidal variation. The complex non-linear features
increases with the increase of Gr values because of higher convec-
tive fluid flow. The interaction of two different fluids (cold and hot)
creates a complex flow pattern as observed in Fig. 4(g). As seen in
the situation, the energy flux vectors are weaker for low Gr values
and getting stronger and showing a clear flow of energy from left
heated portion to the right cold portion through the upper region
of enclosure for increasing Gr values.
Fig. 5 shows the streamlines, isotherms and energy flux vectors
for Gr ¼ 104 , 105 and 106 at Ha ¼ 50, Da ¼ 0:001, / ¼ 0:2 (dashed
lines) when heat source length is L=2. The solid lines in the stream-
lines and isotherms represents the lines corresponding to / ¼ 0:0.
The streamline values are found to be slightly increased at the cen-
tral region of enclosure in Fig. 5 as compared to Fig. 4 for smaller
Grashof number (Gr ¼ 104 and 105 ) but, a clear increase can be
seen when Gr ¼ 106 due to higher heat transfer rate in case of clear
fluid as well as nanofluid. At Gr ¼ 104 , no density variation is found
within the enclosure that predominates a single cellular flow pat-
tern with heated fluid accelerates along the left wall and cold fluid
moves down along the right wall represented in Fig. 5(a). The
streamlines corresponding to nanofluid shows higher deviation
into the opposite direction of heat source as compared to the clear
336 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345

-0.0001
-0.00 004
-0.0
-0.0001

-0.0001

-0.00

02
-0.0001

-0.0003

-0.0001
-0.0003

-0.00
-0.0002
-0.0005
-0.0004
-0.000

-0.0004
-0.0007
-0.0006

-0.0006
-0.0007
-0.0007

-0.000
-0.0003

-0.0009
-0.0004

-0.0008
-0.0008

-0.00
-0.0010
-0.0010

-0.0010
-0.0006
-0.0008

06
07
2

10

8
t=0.05 t=0.10 t=0.15 t=0.20 t=0.25
-0.0001

-0.0001
-0.0002

-0.0001

-0.0002

-0.0001
-0.0002
-0.000

-0.0001

-0.000

-0.0002
-0.0004
-0.00

-0.0005
-0.0002

-0.00

-0.0004
-0.0004

-0.00
-0.000

-0.0006

-0.00

-0.0006
-0.0007
-0.00

-0.000
-0.0008

-0.000
-0.0010

-0.000

-0.0009
-0.0010

-0.0009
4
06

4
06
06
08
08

8
9
t=0.30 t=0.35 t=0.40 t=0.45 t=0.50

0.15
0.2

0.25

0.2
0.25

0.1
0.05

0.3
0.1
0.15

0.15
0.2

0.3

0.35
0.05

0.35
0.15
0.1

0.3
0.35
0.45

0.2
0.05

0.05

0.1
0.45
0.45
0.3

0.45
0.1
0.45

0.2
0.3

0.05
t=0.05 t=0.10 t=0.15 t=0.20 t=0.25
0.25
0.35

0.2
0.3

0.15

0.35
0.35

0.35
0.45

0.35
0.45
0.4
0.45

0.25
0.3
0.4

0.45
0.3

0.25
0.25
0.4
0.45

0.25
0.3
0.4
0.3
0.1

0.05

0.2

0.15
0.2

0.15

0.2

0.15

0.2
0.05
0.05

0.15

0.05
0.1
0.05

0.1

0.1
0.1

t=0.30 t=0.35 t=0.40 t=0.45 t=0.50

Fig. 3. Development of streamlines and isotherms at increasing time level for Gr ¼ 104 , Ha ¼ 10, Da ¼ 0:001 and / ¼ 0:0.

fluid, this effect is almost negligible at Gr ¼ 104 when heat source Da number for 0.001–0.01 the streamline value along the core is
length is L=3. With the increase of Gr, a density inversion is increased from 0.005 to 0.012. With an increase in Darcy number,
observed inside the enclosure resulting a dual cell pattern. At this the cellular pattern shifts to form multicellular and the dual cells
stage, the conductive mode of heat transfer is transferred to con- separates the flow at the core as depicted in Fig. 6(g) and (j). The
vective mode. At Gr ¼ 106 , a pair of center rotating cells appeared flow represents a major conductive cell with a minor inner cell
with maximum density variations and energy transfer along the for small Darcy number and develops a convective regime as the
center of the enclosure due to expansion of anomalous density of heated fluid confines to the boundary layer retarded by the applied
water represented in Fig. 5(g) and (i). The thermal boundary layers magnetic field. The isotherms are almost parallel, straight and hor-
near the heated portion appeared to be larger with larger heat izontal reflecting a convective mode of heat transfer for all Da val-
source length. The isotherms are distributed in vertically upward ues as shown in Fig. 6(b), (e), (h) and (k). The heat transfer is found
direction for small Gr values. As Gr increases the hot fluid moves to be maximum when Da is changed from 0.001 to 0.01. Since the
in upward direction due to buoyancy effect and drives the cold clockwise rotating cell expands and moves away from hot bound-
fluid in opposite direction, which give rise to the convection force. aries. The convective heat transfer can be clearly observed from
energy flux vector plots. The energy flux vectors for Da ¼ 0:001
The central region of the energy flux vectors at Gr ¼ 105 and 106 is
are less dense representing lower convection effects but with the
more dense as compared to the previous case (Fig. 4) where heat
increase in Da values the energy flux vectors becomes stronger
source length is L=3. The energy flux vectors moves from hot to
and forms circular movement of vectors in the core of enclosure.
cold region as the density variation is maximum for L=2.
The energy flux vectors originate from the left active region, pass
In Fig. 6, the steady state isotherms, streamlines and energy flux
through upper region and eliminates from the right cold portion
vectors are presented to show the effect of Darcy number for
of enclosure. For higher Da values, the upper region of the enclo-
Gr ¼ 105 , Ha ¼ 50 and / ¼ 0:20 when length of heat source is sure is mostly occupied by the energy flux vectors, representing
L=3. The intensity of the primary flow is increased due to the the existence of higher convective forces. The comparison of aver-
increase in Da. The low Da number represents a elliptic shape of age heat transfer and entropy generation for the combination of
streamlines along the core of the enclosure for Da ¼ 0:001. Since, various locations of heat source on the vertical walls including
the flow is dominated by the porosity hence with the increase of Middle-Middle, Top-Top, Bottom-Bottom, Bottom-Top and
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345 337

Streamlines Isotherms Energy Flux

-0.00015
-0.0003
-0.0004
-0.00055
-0.00065
-0.0007

-0.0007
0.3 0.25 0.2 0.15
0.45 0.4 0.1
4

0.1 0.05
Gr=10

(a) (b) (c)

-0.001
-0.002
-0.003

-0.0035
-0.004
-0.0
0 45
5
Gr=10

3 0.25
0.
0.3
0.4

0.2
0.2
5
0.1

0.05
0.1

(d) (e) (f)

-0.001
-0.002
-0.004
-0.00
6
-0.0
0 8
-0.009
-0.011 0.350.35
6

0.4
Gr=10

0.3

0 11 0.25
-0 .
0.2
0.05

0.15

0.1

(g) (h) (i)

Fig. 4. Streamline, isotherm and energy flux vector plots at Ha ¼ 50, Da ¼ 0:001 with / ¼ 0:0 (solid lines) and / ¼ 0:20 (dashed lines) for 104 6 Gr 6 106 when heat source
length is L=3.

Top-Bottom combinations for Gr ¼ 106 , Ha ¼ 50, Da ¼ 0:001 and
0:0 6 / 6 0:2 is presented in Fig. 7. The heat transfer is found to
be increasing with nanoparticle volume fractions for all combina-
tions but, in the case Top-Top combination, the increment is lower
as compared to other combinations as shown in Fig. 7(a). The heat
transfer is maximum in case of Bottom-Top combination and is
minimum in Top-Bottom combination. The entropy generation
for different combinations is represented in Fig. 7(b) and is found
to be maximum when both the sources are placed at
Middle-Middle locations and minimum in case of Top-Bottom
combination. The increment in entropy generation with nanoparti-
cle volume fraction is found to be minimum in case of Bottom-
Bottom combination.
The average heat transfer on the left wall is calculated in terms
of average Nusselt number. Total entropy generation by the com-
bined effect of irreversibilities due to heat transfer, fluid friction
and magnetic field is calculated using entropy generation. The
analysis for obtaining the optimum heat transfer on the cost of
least entropy generation is discussed using the ratio of average
Nusselt number with total entropy generation. The dominating
characteristics irreversibility due to heat transfer over irreversibil-
ity because of fluid friction and magnetic field is analyzed by using
Bejan number. The effect of various flow governing parameters on
heat transfer and entropy generation is discussed below.
Fig. 8 shows the local entropy generation due to (a) heat trans-
fer, (b) fluid friction, (c) magnetic field and (d) total entropy gener-
ation due to combined effect of heat transfer, fluid friction and
magnetic field irreversibilities for Gr ¼ 105 , Ha ¼ 10, Da ¼ 0:001
with / ¼ 0:0 (solid lines) and / ¼ 0:2 (dashed lines). The irre-
versibility due to heat transfer is found to be the most dominating
338 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345

Streamlines Isotherms Energy Flux

-0.0002
-0.0004
-0.0005
-0.0006
-0.0008
-0.0008

-0.0008
0.35 0.25

0.4
0.45 0.2 0.15 0.1
4

0.1
Gr=10

0.05

(a) (b) (c)

-0.001
-0.002
-0.003
-0.004

-0.004
5

-0.005
Gr=10

3 0.3
0.
0.3

0.4 0.25
0.2
0.15 0.1 0.05

(d) (e) (f)

-0.002
-0.003
-0.005
-0.007
-0.008
-0.009
-0.011
-0.013 0.35
0.35
6

0.4
Gr=10

-0.014 0.3
0.25
0.2
0.05

0.15
0.1

(g) (h) (i)

Fig. 5. Streamline, isotherm and energy flux vector plots at Ha ¼ 50, Da ¼ 0:001 with / ¼ 0:0 (solid lines) and / ¼ 0:20 (dashed lines) for 104 6 Gr 6 106 when heat source
length L=2.

factor in the entropy generation inside system as compared to the
irreversibility due to fluid friction and magnetic field. Major contri-
bution of heat transfer irreversibility is found in the middle hori-
zontal part of the enclosure and is maximum near the heat
source locations as shown in Fig. 8(a). Minimum contribution of
the local entropy generation in the system is found to be of fluid
friction irreversibility, and maximum values are found near the
side walls of the enclosure as shown in Fig. 8(b). The local entropy
generation due to magnetic field irreversibility is shown in Fig. 8
(c). The irreversibility due to magnetic field mostly occurred at
the lower and upper part of the enclosure. But the contribution
of entropy generation due to magnetic field to the total entropy
generation in the enclosure is much smaller as compared to the
entropy generation due to heat transfer. From Fig. 8(d) it is clear
that total entropy generation inside the enclosure is almost similar
to the entropy generation due to heat transfer which shows that
the contribution of heat transfer entropy generation is maximum
as compared to other factors of entropy generation. Also, it is evi-
dent from Fig. 8(a)–(d) that the entropy generation in case of clear
fluid (/ ¼ 0:0) is less distributed as compared to that of nanofluid
(/ ¼ 0:2), which shows that the entropy generation at higher
nanoparticle volume fraction is more as compared to clear fluid.
Increasing Gr values leads to a considerable change in the heat
transfer and entropy generation. Fig. 9 represents the variation in
average heat transfer, entropy generation and Bejan number for
different values of Grashof number and nanoparticle volume frac-
tion at Ha ¼ 50 with Da ¼ 0:001. Both the volume fraction and Gra-
shof number show a positive impact on the average heat transfer.
From Fig. 9(a), the absolute increase in average Nusselt number is
found to be 0.2128 for Gr ¼ 104 , 0.1196 for Gr ¼ 105 and 0.5313 for
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345 339

Streamlines Isotherms Energy Flux

-0.001

-0.002
-0.003

-0.004
Da=0.001

0.45
-0.005

0.4

5
0.3
0.5
3
0.
25
0.

0.1

0.05
0.2
15
0.

(a) (b) (c)

-0.002
-0.004
-0.006
-0.008
0.4

5
-0.01 0.3
Da=0.01

0.45

-0.0 0.3
12
0.25

0.05
0.2

0.1
5
0.1

(d) (e) (f)

-0.002
-0.004
-0.008
-0.010
-0.01
2
0.4
Da=0.1

-0.0
14
0.3
0.45

-0.0
-0.015

15
0.25
0.05

0.2
0.1

5
0.1

(g) (h) (i)

-0.002
-0.004
-0.006
-0.008
-0.010
-0.01
2
0.4

-0.0
Da=1.0

0.3
5
0.45

14
0.3

-0 -0.0
.0
1 6 15
0.25

0.2
0.05
0.1

5
0.1

(j) (k) (l)

Fig. 6. Streamline, isotherm and energy flux vector plots at Gr ¼ 105 , Ha ¼ 50 and / ¼ 0:20 for 0:001 6 Da 6 1:0.

Gr ¼ 106 on increasing the nanoparticle volume fraction by 20%. ing a combined variation of heat transfer and fluid friction.
The increment of volume fraction in nanofluids depicts an incre- Similarly, from Fig. 9(b), the absolute increase in average entropy
ment of the thermal conductivity and viscosity of nanofluid result- generation in the enclosure is found to be 0.1138, 0.0475, 0.2541
340 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345

2.5

1.5

2
Num

Sm
1.5 1
Middle-Middle
Top-Top Middle-Middle
Bottom-Bottom Top-Top
Bottom-Top Bottom-Bottom
1 Top-Bottom Bottom-Top
Top-Bottom
0.5

0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2

φ φ

Fig. 7. Comparison of (a) average heat transfer and (b) entropy generation with different heat source locations at Gr ¼ 106 , Ha ¼ 50, Da ¼ 0:001 and 0:0 6 / 6 0:2.

1 1
0.00011

0.
2

00
00

00
00

1
0.
0.8 0.8
0.7
1.

0.1836
73

0.6 0.6
0.00040
7.66
091
26

0.3156
0.00001
3.92 7.6662

62

0.00016

0.00011

0.00030
73

3.
0.7091

003
92

0.4 0.4
73

0.00
0.3156
0.1836

0.2 0.2
0.00003

2
00
00
0.00003 0.
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

1 1

0.0030
0.8 0.8
0.0026 0.0570
0.0016
0.0008 1.
93 0.3840
8.202

0.6 0.0002 0.6 31 0.3156

4.

3.3003
17

6
1.9331
09

0.3259
8.20

0.0002
0.4 0.4 0.3840
0.0006
26

0.3156
0.0014
0.0022
0.2 0.0028 0.0 0.2 0.0570
034

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

Fig. 8. (a) Local entropy generation due to heat transfer, (b) local entropy generation due to fluid friction, (c) local entropy generation due to magnetic field and (d) total
entropy generation at Gr ¼ 105 , Ha ¼ 10 Da ¼ 0:01 with / ¼ 0:0 (solid lines) and / ¼ 0:2 (dashed lines).
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345 341

2.5 2

2
1.5
Gr=10 4
Gr=10
5 Gr=10 4
5
Gr=10
6 Gr=10
1.5 Gr=10 6
Num

Sm
1

0.5
0.5

0 0
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
φ φ

2.4
1

2.2

2 0.9
Gr=10 4
5
Gr=10
Gr=10 6
1.8
Num/Sm

Be

0.8
4
1.6 Gr=10
5
Gr=10
Gr=10 6
1.4
0.7

1.2

1 0.6
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
φ φ

Fig. 9. (a) Average heat transfer, (b) entropy generation, (c) PEC (Num =Sm ) and (d) Bejan number with variation of / and Gr at Ha = 50, Da = 0.001.

for Gr ¼ 104 , Gr ¼ 105 and Gr ¼ 106 respectively on 20% increment
of volume fraction. With the change in Grashof number together
with nanoparticle volume fraction we tried to obtain the ratio of
heat transfer to entropy generation. The maximum heat transfer
on the cost of minimum entropy generation is found to be at
Gr ¼ 104 with / ¼ 0:0 when Ha ¼ 50 and Da ¼ 0:001 as given in
Fig. 9(c). Fig. 9(d) represents the Bejan number variation with Gr
and / at Ha ¼ 50 and Da ¼ 0:001 and the value of Bejan number
is found to be greater than 0.5 for all values of Gr and / which
shows the entropy generation dominance due to heat transfer over
entropy generation due to magnetic field and fluid flow. Although
the maximum heat transfer is obtained at higher values of Grashof
number but, at the same time the entropy generation is also at its
maximum for these values. In a recent study Torabi et al. [61] dis-
cussed the performance evaluation criteria which gives the factor
for higher heat transfer rate with minimum entropy generation.
A similar performance evaluation criteria(PEC=Num =Sm ) is used in
the present study for obtaining higher heat transfer rate on the cost
of low entropy generation and is shown in Fig. 9(c), where the
maximum performance is observed at Gr ¼ 104 with / ¼ 0:0.
The average Nusselt number corresponding to varying Ha and /
at Gr ¼ 106 and Da ¼ 0:001 values is shown in 10(a). The effect of
varying Ha is found to be almost negligible on Num at all / values,
but a considerable increase in average Nusselt number is found on
increasing nanoparticle volume fraction. The Num is found to be
maximum at Ha ¼ 1:0 with / ¼ 0:20 with Num ¼ 2:2279. The abso-
lute decrease in Num on increasing Ha from 1 to 50 is 0.0059. But
Num is increased by 0.5313 when / is increased from 0.0 to 0.20,
at Ha ¼ 50. The average entropy generation is found to be maxi-
mum when Ha ¼ 50 and / ¼ 0:20 as shown in Fig. 10(b). The
change in Sm is almost negligible on increasing Ha from 1 to 10,
whereas on increasing the Ha value from 10 to 50 a noticeable
change is found in Sm . Sm is increased by 0.2265 at / ¼ 0:0 and
0.2541 at / ¼ 0:2 when Ha ¼ 50. From Fig. 10(c), maximum heat
transfer at the cost of minimum entropy generation is found at
Ha ¼ 1 with / ¼ 0:20 as the value of Num =Sm is maximum.
342 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345

The values of Num =Sm at Ha ¼ 1 are almost same as that of Ha ¼ 10.
The variation of Bejan number with variation of corresponding Ha
and / values are shown in Fig. 10(d). Be values are greater than 0.5
for all Ha and /, which signifies that irreversibility due to heat
transfer is dominant over irreversibility due to magnetic field
and fluid flow. Be is maximum when Ha ¼ 1, hence the entropy
generation due to heat transfer is maximum when Ha ¼ 1. It can
be observed from Figs. 10(c) and (d), the Num =Sm values at
Ha ¼ 1 and 10 are almost same but a noticeable difference is found
in Be values at Ha ¼ 1 and 10. Hence, it can be concluded that heat
transfer entropy generation is more when Ha ¼ 1 as compared to
Ha ¼ 10 (from Fig. 10(d)), whereas the total entropy generation
is almost same for these values. Finally, from Fig. 10(c) it can be
concluded that maximum performance of system can be obtained
at Ha ¼ 1 with / ¼ 0:20 when Gr ¼ 106 and Da ¼ 0:001, as the
higher values of Ha leads to higher entropy generation in the sys-
tem and decreases the performance level.
Fig. 11(a) shows the variation of Num with corresponding Da
and / values. The effect of increasing Da is clearly seen on Num
when Da values are increased from 0.001 to 0.1, but the increase
in Num becomes almost negligible when Da is increased from 0.1
to 1.0. Also, it is clear from the figure that the difference between
Num values decreases for higher Da values. But, when the entropy
generation is considered, the higher values of Da generates higher
entropy as shown in Fig. 11(b). The difference between Sm is clearly
visible even for higher values of Da. The absolute increase in Num
on increasing Da from 0.001 to 1 is found to be 0.3872, 0.4692,
0.5621 and 0.7873 for / = 0.0, 0.05, 0.1 and 0.2 respectively. Simi-
larly, absolute increase in Sm on increasing Da from 0.001 to 1 is
found to be 0.4064, 0.4867, 0.5805 and 0.8227 for / = 0.0, 0.05,
0.1 and 0.2 respectively. The optimum value for maximum heat
transfer and minimum entropy generation is at obtained for
Da ¼ 0:001 and / ¼ 0:2 for Gr ¼ 106 and Ha ¼ 50, where Num =Sm
is 1.3286 as shown in Fig. 11(c). Fig. 11(d) shows the variation of
Be with Da and / at Gr ¼ 106 and Ha ¼ 50. The heat transfer irre-
versibility is dominant over irreversibility due to magnetic field
and fluid friction for Da P 0:1 for all values of / except when
/ ¼ 0:2. When Da ¼ 1, the heat transfer irreversibility is dominant
for clear fluid, but with the involvement of nanoparticles in the
fluid the combined effect of magnetic field and fluid friction

1.7
2.2

2.1

1.6
2
Num

Sm

1.9
1.5
Ha=1 Ha=1
1.8 Ha=10 Ha=10
Ha=50 Ha=50

1.7
1.4
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
φ φ

1.4 1

1.35

0.9

1.3
Num/Sm

Be

0.8
1.25

Ha=1
Ha=10
1.2 Ha=1
Ha=50
Ha=10 0.7
Ha=50

1.15
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
φ φ

Fig. 10. (a) Average heat transfer, (b) entropy generation, (c) PEC (Num =Sm ) and (d) Bejan number with variation of / and Ha at Gr ¼ 106 , Da ¼ 0:001.
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345 343

2.4
2.8

2.6 2.2

2.4
2

Sm
Num

Da=0.001
2.2 Da=0.01
1.8 Da=0.1
Da=1.0
2
Da=0.001
Da=0.01 1.6
1.8 Da=0.1
Da=1.0

1.6 1.4
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
φ φ

1.35 0.7

0.65
1.3
Da=0.001
0.6 Da=0.01
Da=0.1
1.25 Da=1.0
Nu m/Sm

Be

0.55

1.2
0.5

Da=0.001
1.15 Da=0.01
0.45
Da=0.1
Da=1.0

1.1 0.4
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
φ φ

Fig. 11. (a) Average heat transfer, (b) entropy generation, (c) PEC (Num =Sm ) and (d) Bejan number with variation of / and Da at Gr ¼ 106 , Ha ¼ 50.

irreversibility is dominant over heat transfer irreversibility. The
least value of Be is found for / ¼ 0:2 with Da ¼ 1.
5. Conclusion
Heat transfer and entropy generation inside a closed chamber is
encountered in various industrial applications, such as geothermal
energy extraction, nuclear waste storage, solidification process of
metals and alloys. A finite volume simulation has been carried out
for the numerical study of heat transfer and entropy generation of
a MHD flow inside a porous medium. Time periodic boundary con-
ditions are used for temperature distribution on the mid portion of
left wall, the adjacent side wall is maintained at comparatively cold
temperature. Time history analysis has been presented for
Gr ¼ 104 , Ha ¼ 10, Da ¼ 0:001 and / ¼ 0:0. Further, for the discus-
sion of PEC the results are calculated for 104 6 Gr 6 106 ,
1 6 Ha 6 50, 0:001 6 Da 6 1 and 0:0 6 / 6 0:2. The important
results of the conducted study have been listed as follows:
 The flow is dominated by convection effects when Gr values
shifts from 104 to 105 . For higher values of Grashof number
the fluid is well mixed and more entropy is generated.
 Higher heat transfer and entropy generation is found for higher
Grashof number. But the ratio of average Nusselt number to
average entropy is large when Gr ¼ 104 , which shows the
higher heat transfer on the cost of least entropy generation.
 Based on the results obtained, it can be concluded that for
Gr ¼ 106 and Ha ¼ 50, the maximum heat transfer on the cost
of least entropy can be obtained when Da ¼ 0:001 with / ¼ 0:2.
 It is found that when Gr ¼ 106 and Da ¼ 0:001, the rate of heat
transfer almost constant for varying Ha, but the entropy gener-
ation is higher when Ha is increased from 10 to 50. The heat
transfer irreversibility is found to be dominant over magnetic
field and fluid friction irreversibility. The total entropy genera-
tion is almost equal at Ha ¼ 1 and 10, but the entropy genera-
tion due to heat transfer is more when Ha ¼ 1 as compared to
Ha ¼ 10.
344 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345

 The average heat transfer rate increases with increase in Da, but [22] S.J. Kim, K. Vafai, Analysis of natural convection about a vertical plate
embedded in a porous medium, Int. J. Heat Mass Transfer 32 (4) (1989) 665–
it become almost constant for Da > 0:1. Total entropy genera-
677.
tion in the system increases with increase in Da. The heat trans- [23] A.A. Mohamad, Natural convection from a vertical plate in a saturated porous
fer irreversibility is dominant over irreversibility due to medium: nonequilibrium theory, J. Porous Media 4 (2) (2001) 181–186.
magnetic field and fluid friction when Da > 0:1. For Da ¼ 0:1, [24] O.M. Haddad, M.A. Al-Nimr, A.N. Al-Khateeb, Validation of the local thermal
equilibrium assumption in natural convection from a vertical plate embedded
the irreversibility due to heat transfer is dominant for / 6 0:1 in porous medium: non-Darcian model, Int. J. Heat Mass Transfer 47 (8-9)
and for Da ¼ 1 heat transfer irreversibility is found to be dom- (2004) 2037–2042.
inant for clear fluid (i.e. / ¼ 0:0). [25] Abdalla Mohamed Al-Amiri, Natural convection in porous enclosures: the
application of the two-energy equation model, Numer. Heat Transfer: Part A:
 The fractional increment of Cu nanoparticles in water increases Appl. 41 (8) (2002) 817–834.
the heat transfer rate very effectively. [26] Feng Wu, Gang Wang, Wenjing Zhou, Natural convection in a cavity filled with
porous medium with partially thermal active sidewalls under local thermal
nonequilibrium conditions, J. Porous Media 17 (11) (2014).
[27] Tiew Wei Ting, Yew Mun Hung, Ningqun Guo, Viscous dissipative forced
convection in thermal non-equilibrium nanofluid-saturated porous media
embedded in microchannels, Int. Commun. Heat Mass Transfer 57 (2014) 309–
References 318.
[28] Tiew Wei Ting, Yew Mun Hung, Ningqun Guo, Entropy generation of viscous
dissipative nanofluid flow in thermal non-equilibrium porous media
[1] Mikhail A. Sheremet, Ioan Pop, Natalia C. Roca, Magnetic field effect on the
embedded in microchannels, Int. J. Heat Mass Transfer 81 (2015) 862–877.
unsteady natural convection in a wavy-walled cavity filled with a nanofluid:
[29] T. Armaghani, Ali J. Chamkha, Mahmoud Maghrebi, Mohsen Nazari, Numerical
Buongiorno’s mathematical model, J. Taiw. Inst. Chem. Eng. 61 (2016) 211–
analysis of a nanofluid forced convection in a porous channel: a new heat flux
222.
model in LTNE condition, J. Porous Media 17 (7) (2014).
[2] Nadezhda S. Bondareva, Mikhail A. Sheremet, Hakan F. Oztop, Nidal Abu-
[30] Guillermo Ibáñez, Entropy generation in MHD porous channel with
Hamdeh, Heatline visualization of MHD natural convection in an inclined
hydrodynamic slip and convective boundary conditions, Int. J. Heat Mass
wavy open porous cavity filled with a nanofluid with a local heater, Int. J. Heat
Transfer 80 (2015) 274–280.
Mass Transfer 99 (2016) 872–881.
[31] Sadegh Khalili, Saeed Dinarvand, Reza Hosseini, Hossein Tamim, Ioan Pop,
[3] Sadegh Khalili, Hossein Tamim, Arezoo Khalili, Mohammad Mehdi Rashidi,
Unsteady MHD flow and heat transfer near stagnation point over a stretching/
Unsteady convective heat and mass transfer in pseudoplastic nanofluid over a
shrinking sheet in porous medium filled with a nanofluid, Chin. Phys. B 23 (4)
stretching wall, Adv. Powder Technol. 26 (5) (2015) 1319–1326.
(2014) 048203.
[4] I. Fersadou, H. Kahalerras, M. El Ganaoui, MHD mixed convection and entropy
[32] GHR Kefayati, Heat transfer and entropy generation of natural convection on
generation of a nanofluid in a vertical porous channel, Comp. Fluids 121 (2015)
non-newtonian nanofluids in a porous cavity, Powder Technol. 299 (2016)
164–179.
127–149.
[5] N. Rudraiah, R.M. Barron, M. Venkatachalappa, C.K. Subbaraya, Effect of a
[33] G.H.R. Kefayati, Nor Azwadi Che Sidik, Simulation of natural convection and
magnetic field on free convection in a rectangular enclosure, Int. J. Eng. Sci. 33
entropy generation of non-newtonian nanofluid in an inclined cavity using
(8) (1995) 1075–1084.
buongiorno’s mathematical model (part ii, entropy generation), Powder
[6] Hakan F. Oztop, M.M. Rahman, A. Ahsan, M. Hasanuzzaman, R. Saidur, Khaled
Technol. 305 (2017) 679–703.
Al-Salem, N.A. Rahim, MHD natural convection in an enclosure from two semi-
[34] I. Fersadou, H. Kahalerras, M. El Ganaoui, MHD mixed convection and entropy
circular heaters on the bottom wall, Int. J. Heat Mass Transfer 55 (7–8) (2012)
generation of a nanofluid in a vertical porous channel, Comp. Fluids 121 (2015)
1844–1854.
164–179.
[7] S. Sivasankaran, V. Sivakumar, P. Prakash, Numerical study on mixed
[35] S. Hussain, K. Mehmood, M. Sagheer, MHD mixed convection and entropy
convection in a lid-driven cavity with non-uniform heating on both
generation of wateralumina nanofluid flow in a double lid driven cavity with
sidewalls, Int. J. Heat Mass Transfer 53 (19–20) (2010) 4304–4315.
discrete heating, J. Magnet. Magnet. Mater. 419 (2016) 140–155.
[8] S. Sivasankaran, A. Malleswaran, Jinho Lee, Pon Sundar, Hydro-magnetic
[36] Adrian Bejan, Convection Heat Transfer, John Wiley & Sons, 2013.
combined convection in a lid-driven cavity with sinusoidal boundary
[37] Ram Satish Kaluri, Tanmay Basak, Heatline analysis of thermal mixing due to
conditions on both sidewalls, Int. J. Heat Mass Transfer 54 (1–3) (2011) 512–
natural convection in discretely heated porous cavities filled with various
525.
fluids, Chem. Eng. Sci. 65 (6) (2010) 2132–2152.
[9] S. Sivasankaran, K.L. Pan, Numerical simulation on mixed convection in a
[38] Tanmay Basak, Ali J. Chamkha, Heatline analysis on natural convection for
porous lid-driven cavity with nonuniform heating on both side walls, Numer.
nanofluids confined within square cavities with various thermal boundary
Heat Transf., Part A: Appl. 61 (2) (2012) 101–121.
conditions, Int. J. Heat Mass Transfer 55 (21–22) (2012) 5526–5543.
[10] S.U.S. Chol et al., Enhancing thermal conductivity of fluids with nanoparticles,
[39] Kamel Hooman, Energy flux vectors as a new tool for convection visualization,
ASME-Publications-Fed 231 (1995) 99–106.
Int. J. Numer. Meth. Heat Fluid Flow 20 (2) (2010) 240–249.
[11] Jeffrey A. Eastman, SUS Choi, Sheng Li, W. Yu, L.J. Thompson, Anomalously
[40] Pallab S. Mahapatra, Nirmal K. Manna, Koushik Ghosh, Achintya
increased effective thermal conductivities of ethylene glycol-based nanofluids
Mukhopadhyay, Heat transfer assessment of an alternately active bi-heater
containing copper nanoparticles, Appl. Phys. Lett. 78 (6) (2001) 718–720.
undergoing transient natural convection, Int. J. Heat Mass Transfer 83 (2015)
[12] Hadi Ghodsinezhad, Mohsen Sharifpur, Josua Petrus Meyer, Experimental
450–464.
investigation on cavity flow natural convection of Al2O3–water nanofluids, Int.
[41] N. Nithyadevi, P. Kandaswamy, S. Malliga Sundari, Magnetoconvection in a
Commun. Heat Mass Transfer 76 (2016) 316–324.
square cavity with partially active vertical walls: time periodic boundary
[13] Sumit Malik, A.K. Nayak, A comparative study of mixed convection and its
condition, Int. J. Heat Mass Transfer 52 (7) (2009) 1945–1953.
effect on partially active thermal zones in a two sided lid-driven cavity filled
[42] A.K. Nayak, S. Malik, K. Venkateshwarlu, P.K. Jena, Magneto-convection and its
with nanofluid, Eng. Sci. Technol., Int. J. (2016), pages –.
effect on partially active thermal zones in a porous square domain, Int. J. Heat
[14] Mohsen Sheikholeslami, Mofid Gorji Bandpy, R. Ellahi, Ahmad Zeeshan,
Mass Transfer 95 (2016) 913–926.
Simulation of MHD CuO–water nanofluid flow and convective heat transfer
[43] Zouhaier Mehrez, Afif El Cafsi, Ali Belghith, Patrick Le Qur, {MHD} effects on
considering lorentz forces, J. Magnet. Magnet. Mater. 369 (2014) 69–80.
heat transfer and entropy generation of nanofluid flow in an open cavity, J.
[15] B. Ghasemi, S.M. Aminossadati, A. Raisi, Magnetic field effect on natural
Magnet. Magnet. Mater. 374 (2015) 214–224.
convection in a nanofluid-filled square enclosure, Int. J. Thermal Sci. 50 (9)
[44] Sumit Malik, A.K. Nayak, Buoyancy driven heat transfer in nanofluids due to
(2011) 1748–1756.
wall mounted heat source, Alex. Eng. J. 55 (2) (2016) 797–810.
[16] Guillermo Ibáñez, Aracely López, Joel Pantoja, Joel Moreira, Entropy generation
[45] G.H.R. Kefayati, Simulation of heat transfer and entropy generation of {MHD}
analysis of a nanofluid flow in MHD porous microchannel with hydrodynamic
natural convection of non-newtonian nanofluid in an enclosure, Int. J. Heat
slip and thermal radiation, Int. J. Heat Mass Transfer 100 (2016) 89–97.
Mass Transfer 92 (2016) 1066–1089.
[17] Mohammad Hassan Saidi, Hossein Tamim, Heat transfer and pressure drop
[46] M.A. Mansour, M.A.Y. Bakier, Influence of thermal boundary conditions on
characteristics of nanofluid in unsteady squeezing flow between rotating
MHD natural convection in square enclosure using Cu–water nanofluid,
porous disks considering the effects of thermophoresis and brownian motion,
Energy Reports 1 (2015) 134–144.
Adv. Powder Technol. 27 (2) (2016) 564–574.
[47] Ali Al-Zamily, M. Ruhul Amin, Natural convection and entropy generation in a
[18] Hillal M. Elshehabey, Sameh E. Ahmed, MHD mixed convection in a lid-driven
cavity filled with two horizontal layers of nanofluid and porous medium in
cavity filled by a nanofluid with sinusoidal temperature distribution on the
presence of a magnetic field, in: Proceedings of the ASME 2015 International
both vertical walls using Buongiornos nanofluid model, Int. J. Heat Mass
Mechanical Engineering Congress and Exposition, vol. 8B, 2015, pp. 913–926
Transfer 88 (2015) 181–202.
[48] R.K. Nayak, S. Bhattacharyya, I. Pop, Numerical study on mixed convection and
[19] B. Pekmen, M. Tezer-Sezgin, MHD flow and heat transfer in a lid-driven porous
entropy generation of Cu–water nanofluid in a differentially heated skewed
enclosure, Comp. Fluids 89 (2014) 191–199.
enclosure, Int. J. Heat Mass Transfer 85 (2015) 620–634.
[20] M.A. Al-Nimr, O.M. Haddad, Fully developed free convection in open-ended
[49] GH. R. Kefayati, Mesoscopic simulation of mixed convection on non-
vertical channels partially filled with porous material, J. Porous Media 2 (2)
newtonian nanofluids in a two sided lid-driven enclosure, Adv. Powder
(1999) 179–189.
Technol. 26 (2) (2015) 576–588.
[21] D.AndrewS. Rees, Darcy-Brinkman free convection from a heated horizontal
surface, Numer. Heat Transfer: Part A: Appl. 35 (2) (1999) 191–204.
 S. Malik, A.K. Nayak / International Journal of Heat and Mass Transfer 111 (2017) 329–345
[50] GH.R. Kefayati, {FDLBM} simulation of magnetic field effect on mixed
convection in a two sided lid-driven cavity filled with non-newtonian
nanofluid, Powder Technol. 280 (2015) 135–153.
[51] James Clerk Maxwell, A Treatise on Electricity and Magnetism, vol. 1,
Clarendon press, 1881.
[52] Ali Meerali Jasim Al-Zamily, Analysis of natural convection and entropy
generation in a cavity filled with multi-layers of porous medium and nanofluid
with a heat generation, Int. J. Heat Mass Transfer 106 (2017) 1218–1231.
[53] Pratibha Biswal, Avijit Nag, Tanmay Basak, Analysis of thermal management
during natural convection within porous tilted square cavities via heatline and
entropy generation, Int. J. Mech. Sci. 115116 (2016) 596–615.
[54] M.A. Mansour, Sameh E. Ahmed, A numerical study on natural convection in
porous media-filled an inclined triangular enclosure with heat sources using
nanofluid in the presence of heat generation effect, Eng. Sci. Technol., Int. J. 18
(3) (2015) 485–495.
[55] Yasin Varol, Hakan F. Oztop, Ioan Pop, Entropy generation due to natural
convection in non-uniformly heated porous isosceles triangular enclosures at
different positions, Int. J. Heat Mass Transfer 52 (5) (2009) 1193–1205.
345
[56] B. Hoseinpour, Hamid Reza Ashorynejad, Koroush Javaherdeh, Entropy
generation of nanofluid in a porous cavity by lattice Boltzmann method, J.
Thermophys. Heat Transfer 1–8 (2016).
[57] Tanmay Basak, Ram Satish Kaluri, A.R. Balakrishnan, Entropy generation
during natural convection in a porous cavity: effect of thermal boundary
conditions, Numer. Heat Transf., Part A: Appl. 62 (4) (2012) 336–364.
[58] S. Bhattacharyya, A.K. Singh, Vortex shedding and heat transfer dependence on
effective reynolds number for mixed convection around a cylinder in cross
flow, Int. J. Heat Mass Transfer 53 (1516) (2010) 3202–3212.
[59] M. Sathiyamoorthy, Tanmay Basak, S. Roy, I. Pop, Steady natural convection
flow in a square cavity filled with a porous medium for linearly heated side
wall(s), Int. J. Heat Mass Transfer 50 (9-10) (2007) 1892–1901.
[60] C.J. Ho, W.K. Liu, Y.S. Chang, C.C. Lin, Natural convection heat transfer of
alumina-water nanofluid in vertical square enclosures: an experimental study,
Int. J. Thermal Sci. 49 (8) (2010) 1345–1353.
[61] Mehrdad Torabi, G.P. Peterson, Mohsen Torabi, Nader Karimi, A
thermodynamic analysis of forced convection through porous media using
pore scale modeling, Int. J. Heat Mass Transfer 99 (2016) 303–316.