Alexandria Engineering Journal (2016) 55, 127–139

H O S T E D BY
Alexandria University

Alexandria Engineering Journal

www.elsevier.com/locate/aej
www.sciencedirect.com

ORIGINAL ARTICLE

Numerical investigation of pure mixed convection

in a ferrofluid-filled lid-driven cavity for different
heater configurations
Khan Md. Rabbi a, Sourav Saha a, Satyajit Mojumder a, M.M. Rahman b,*,
R. Saidur c, Talaat A. Ibrahim d,e

a
Department of Mechanical Engineering, Bangladesh University of Engineering and Technology, Dhaka 1000, Bangladesh
b
Universiti Brunei Darussalam, Mathematical and Computing Sciences Group, Faculty of Science, Brunei
c
Centre of Research Excellence in Renewable Energy, King Fahd University of Petroleum and Minerals, Dhahran 31261,
Saudi Arabia
d
Vice Presidency for Projects, King Saud University, P.O. 70908, 11577 Riyadh, Saudi Arabia
e
Mechanical Power Dept., Faculty of Engineering-Mattaria, Helwan University, Cairo 11718, Egypt

Received 3 August 2015; revised 8 December 2015; accepted 20 December 2015

Available online 12 January 2016

KEYWORDS Abstract Mixed convection has been a center point of attraction to the heat transfer engineers for
Mixed convection; many years. Here, pure mixed convection analysis in cavity is carried out for two different geomet-
Ferrofluid; ric heater configurations under externally applied magnetic field. Ferrofluid (Fe3O4–water) is con-
Lid-driven cavity; sidered as working fluid and modeled as single phase fluid. The heaters at the bottom wall are kept
MHD convection; at constant high temperature while vertical side walls are adiabatic. The top wall is moving at a con-
Bottom heater stant velocity in both geometric configurations and is kept at constant low temperature. Galerkin
weighted residuals method of finite element analysis is implemented to solve the governing equa-
tions. The analysis has been carried out for a wide range of Richardson number (Ri = 0.1–10), Rey-
nolds number (Re = 100–500), Hartmann number (Ha = 0–100) and solid volume fraction
(u = 0–0.15) of ferrofluid. The overall heat transfer performance for both the configurations is
quantitatively investigated by average Nusselt number at the heated boundary wall. It is observed
that higher Ri enhances the heat transfer rate, although higher Ha decreases heat transfer rate.
Moreover, at higher Ri and lower Ha, semi-circular notched cavity shows significantly better (more
than 30%) heat transfer rate.
Ó 2016 Faculty of Engineering, Alexandria University. Production and hosting by Elsevier B.V. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
* Corresponding author. Tel.: +673 2463001x1355; fax: +673
2461502. Magneto-hydrodynamic (MHD) convection is the heat trans-
E-mail address: mustafizur.rahman@ubd.edu.bn (M.M. Rahman).
fer mechanism in fluid under external magnetic field. This topic
Peer review under responsibility of Faculty of Engineering, Alexandria
has turned out to be a focal point of enormous amount of
University.
http://dx.doi.org/10.1016/j.aej.2015.12.021
1110-0168 Ó 2016 Faculty of Engineering, Alexandria University. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
128
Nomenclature
a notch spacing length (m)
g acceleration due to gravity (m s
2)
k thermal conductivity of fluid (W m
1 K
1)
L length of the square cavity (m)
Nu average Nusselt number on heater surface
p pressure (Pa)
P dimensionless pressure
Pr Prandtl number
Ha Hartmann number
Re Reynolds number
Gr Grashof number
Ri Richardson number
s heater wall perimeter
T temperature (K)
u velocity at x-direction (m s
1)
U dimensionless velocity at x-direction
v velocity at y-direction (m s
1)
V dimensionless velocity at y-direction
X dimensionless distance along x-coordinate
Y dimensionless distance along y-coordinate
researches from the last three decades because of its wide range
of application in petroleum industries, plasma physics, geo-
physics, MHD pump, MHD flow meter, cooling of nuclear
reactors and various thermal systems [1–3].
Previously, several investigations on natural, forced and
mixed convection have been performed under external mag-
netic field. It was reported by many researchers that natural
convection can be affected by external magnetic field [4–7].
Rudraiah et al. [8] investigated the effect of magnetic field on
natural convection in a rectangular enclosure numerically
and presented a correlation formula of Nusselt number. They
found that the effect of the magnetic field decreased the rate of
heat transfer at any Grashof number. Similar studies have
been performed for different geometries such as square enclo-
sure [9,10], semi-annulus enclosure [11,12], and trapezoidal
enclosure [13]. It is also noticeable that many researchers inves-
tigated MHD natural convection using nanoparticles impro-
vised in a base fluid [14–20]. Mahmoudia et al. [21]
investigated MHD natural convection and entropy generation
in a trapezoidal enclosure using Cu–water nanofluid. They
found that enhancement of Nusselt number occurred with
increasing Hartmann number at Ra = 104 and 105, but at even
higher Rayleigh number, opposite occurred. Moreover, fer-
rofluid also has become very popular while achieving higher
heat transfer rate [22,23]. Recently, few researchers investi-
gated natural convection in different geometric configurations
using ferrofluids [24–26]. Gavili et al. [27] investigated thermo-
magnetic convection process in two-dimensional enclosure
filled with ferrofluid. It was concluded that average Nusselt
number increased when magnetic field and temperature gradi-
ent had same direction. Lajvardi et al. [28] carried out experi-
mental study on natural convection using ferrofluid under
external magnetic field. It was reported that significant
enhancement of heat transfer was present at lower strength
of magnetic field. Bondareva et al. [29] investigated the effect
K.Md. Rabbi et al.
Greek symbols
a thermal diffusivity (m2 s
1)
b volume expansion coefficient (K
1)
u solid volume fraction of ferrofluid
l dynamic viscosity (Ns m
2)
m kinematic viscosity (m2 s
1)
q density (kg m
3)
r electrical conductivity ((X m)
1)
w stream function
H dimensionless temperature
Subscripts
h hot
c cold
f base fluid (water)
ff ferrofluid
p ferrous particle
of uniform magnetic field on laminar flow regimes of natural
convection in an enclosure. It was observed that Rayleigh
number, Hartmann number and orientation of magnetic field
had impact on velocity distribution, temperature field and
average Nusselt number.
Many researchers tried to figure out how external magnetic
field affected heat transfer rate while analyzing mixed convec-
tion [30–32]. Oztop et al. [33] studied MHD mixed convection
in an open channel with a fully or partially heated cavity. It
was evident that higher value of Ha did not lead to any change
in flow field for any length of heater. It was also reported that
at any values of Richardson number, higher amount of
nanoparticles increased heat transfer rate. Some related inves-
tigations for different cavities and boundary conditions can be
found in [34–36]. Many researchers reported that nanofluids
and ferrofluids had significant influence on mixed convection
[37,38]. Ahmed et al. [39] investigated MHD mixed convection
of a lid-driven cavity filled with nanofluid. It was investigated
that the existence of an inclined magnetic field decreased the
fluid movement and hence decreased the heat transfer rate.
Sheremet et al. [40] studied mixed convection in a lid-driven
square cavity with nanofluid inside. It was evident that
Prandtl, Reynolds number, Grashof number, Lewis number,
buoyancy-ratio, the Brownian motion, and the thermophoresis
had effect on flow and heat transfer. Aminfar et al. [41] inves-
tigated numerically the hydro-thermal characteristics of a fer-
rofluid in a vertical rectangular duct which was exposed to a
non-uniform transverse magnetic field.
The present study is concerned with the comparison of heat
transfer performance between semi-circular and triangular
notches in lid-driven cavities. The purpose of this investigation
is to study heater geometry effect on pure mixed convection
heat transfer. Ferrofluid has been considered as the working
fluid. The flow and thermal field have been studied and dis-
cussed by streamline and isothermal contours respectively.
Pure mixed convection in a ferrofluid-filled lid-driven cavity 129

uo uo

a)
(a) b)(b)

Figure 1 Schematic diagram of (a) triangular and (b) semi-circular notched cavity.

2D plots are included to show the effect of different pertinent nesq approximation is assumed to hold true for buoyancy
parameters such as Hartmann number (Ha), Reynolds number effect. Under these assumptions, the governing equations for
(Re), Grashof number (Gr), and Richardson number (Ri) for the steady, two-dimensional, laminar and incompressible
different heater geometries. Finally the consequence of varying mixed convection flow are expressed as follows:
solid volume fraction (u) of ferrofluid has been illustrated and @u @v
discussed. þ ¼ 0; ð1Þ
@x @y

2 
2. Problem formulation @u @v 1 @p @ u @2 u
u þv ¼
þ mff þ ; ð2Þ
@x @y qff @x @x2 @y2
2.1. Physical model
2 
@v @v 1 @p @ v @2 v rff
u þv ¼
þ mff þ 2 þ bff gðT
Tc Þ
B2o v; ð3Þ
@x @y qff @y @x @y
2 qff
The geometries of two models are shown in Fig. 1 with appro-

2 
priate dimensions and boundary conditions. The models con- @H @H @ H @2H
u þv ¼ aff þ : ð4Þ
sidered here are two dimensional square enclosures with @x @y @x2 @y2
sides of length L and non-dimensional notch spacing of
The effective thermo-physical properties of ferrofluids are
a/L = 0.5. Both enclosures, one with triangular notch and
defined by using the following formula:
the other with semi-circular notch, are filled with Fe3O4–water
The effective density of ferrofluid is obtained by,
ferrofluid. The horizontal top walls are moving with a constant
velocity of Uo in both the cases. Both the notches (triangular qff ¼ ð1
uÞqf þ uqp : ð5Þ
and semi-circular) are kept at high temperature of Th whereas
the opposite top walls are kept at low temperature of Tc. The Under the thermal equilibrium conditions, the specific heat
left and right side walls and also the bottom walls except notch of ferrofluid is given as,
surfaces are kept insulated in both enclosures. Moreover, a
ðqCp Þff ¼ ð1
uÞðqCp Þf þ uðqCp Þp : ð6Þ
uniform magnetic field is applied horizontally normal to the
vertical wall. Gravity is assumed to act along the negative Electrical conductivity of ferrofluid [43] is given as follows
y-direction. The radiation, pressure work and viscous dissipa-
rff ¼ ð1
uÞrf þ urp : ð7Þ
tion are assumed to be negligible. Fe3O4 nanoparticles are
assumed to have a uniform shape and size. It is also assumed Thermal expansion coefficient of ferrofluid is given by,
that the nanoparticles and water are in thermal equilibrium,
ðbqÞff ¼ ð1
uÞðbqÞbf þ uðqbÞp : ð8Þ
and the ferrofluid is Newtonian and incompressible. For math-
ematical modeling, the fluid is assumed to be of single phase The dynamic viscosity of ferrofluid is expressed by Brink-
continuum. The flow is considered to be steady, two dimen- man model [44],
sional and laminar. The displacement currents, induced mag-
netic field, viscous dissipation and radiation heat transfer are lff ¼ lf ð1
uÞ
0:25 : ð9Þ
also neglected. The thermal conductivity of the ferrofluid is obtained by
Maxwell-Garnett model [45] and stated as
2.2. Mathematical modeling  
ðkp þ 2kf Þ
2uðkf
kp Þ
kff ¼ kf : ð10Þ
2.2.1. Governing equation ðkp þ 2kf Þ þ uðkf
kp Þ
The governing equations of the problem specification are con- Eqs. (1)–(4) can be converted to the dimensionless form by
servation of mass, momentum and energy equations. Boussi- using the following dimensionless parameters:
130 K.Md. Rabbi et al.

Table 1 Present boundary condition in the nondimensional Table 2 Thermo-physical properties of Fe3O4 (nanoparticles)
form. and water (base fluid) [42].
Boundary Flow field Thermal field Property Fe3O4 Water (base
(nanoparticles) fluid)
Top wall U = 1, V = 0 H¼0
@H
@X ¼ 0
Left vertical wall U=V=0 Heat capacitance 670 4179
@H (J Kg
1 K
1)
@X ¼ 0
Right vertical wall U=V=0
@H Density (Kg m
3) 5200 997.1
@Y ¼ 0
Bottom wall U=V=0
0 6 X 6 0:25; 0:75 6 X 6 1 Thermal conductivity 6 0.613
Triangular notched wall U=V=0 H¼1 (Wm
1 K
1)
Semi-circular notched wall U=V=0 H¼1 Thermal expansion coefficient 1.18  10
5 2.1  10
4
(K
1)
Dynamic viscosity (Ns m
2) – 0.001003

2.2.2. Boundary conditions

x y u v Boundary conditions illustrated in Fig. 1 can be written in
X¼ ; Y¼ ; U¼ ; V¼ ;
L L u0 u0 dimensionless forms and are presented in Table 1. Both flow
p T
Tc and thermal boundary conditions have been shown in tabular
P¼ ; H¼ : ð11Þ
qnp u2o Th
Tc form at top wall, vertical walls, bottom wall and notch (heater)
wall.
gbf ðTh
Tc ÞL3 mf
Gr ¼ ; Pr ¼ ;
m2f af 2.3. Properties of ferrofluid
rffiffiffiffiffiffiffiffiffi
rff Uo L Gr
Ha ¼ B0 H ; Re ¼ ; Ri ¼ 2 : ð12Þ
qff mf mf Re As described in the introduction section, the present investiga-
Therefore using the above parameters leads to dimension- tion has considered ferrofluid (Fe3O4–water) as the working
less forms of the governing equations as below: fluid. The thermo-physical properties of nanoparticle (Fe3O4)
and base fluid (water) have been summarized in Table 2. These
@U @V properties have been considered constant except for density
þ ¼ 0; ð13Þ
@X @Y variation according to Boussinesq approximation.


@U @U @P 1 mff @ 2 U @ 2 U
U þV ¼
þ þ 2 ; ð14Þ
@X @Y @X Re mf @X @Y2
3. Numerical procedure


@V @V @P 1 mff @ 2 V @ 2 V bff Ha2
U þV ¼
þ þ 2 þ RiH
V: ð15Þ
@X @Y @Y Re mf @X @Y
2
bf Re The entire domain is fragmented into non-overlapping trian-

2 2 
@H @H 1 aff @ H @ H gular mesh elements of different sizes. To solve the system
U þV ¼ þ : ð16Þ
@X @Y RePr af @X2 @Y2 along with specified boundary conditions, Galerkin weighted
residual method of finite element analysis has been employed.
Local Nusselt number can be expressed as, Non-dimensional continuity, momentum and energy equations
kff @H are converted into set of algebraic equations. P2–P1 Lagrange
Nu ¼
: ð17Þ
kf @n finite elements are used to discretize pressure and velocity com-
ponents and Lagrange-quadratic finite elements are chosen for
and temperature. Iterative procedure is used to find the solution
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2
2 until the convergence occurs. The criterion of convergence is
@H @H @H
¼ þ : ð18Þ set to 10
6, so that jzpþ1
zp j 6 10
6 , where z is the general
@n @X @Y
dependent variable and p is the number of iteration. When
The characteristics of heat transfer are obtained by average the difference exceeds the criterion, iteration terminates.
Nusselt number and can be expressed as
Z 3.1. Grid independency test
1 s
Nu ¼ Nuds: ð19Þ
s 0
A grid independency test has been carried out to find numeri-
Here, s signifies the circumference. For triangular notch, cal accuracy while solving the system. The test is done for
L Gr = 104, Ri = 0.1, Ha = 25 and u = 0.15. Fig. 2 illustrates
s ¼ pffiffiffi ð20Þ the effect of mesh elements on mid-plane velocity profile (U)
2
for both triangular and semi-circular notched cavities. From
and for semi-circular notch, Fig. 2(a) it is evident that the velocity profile almost coincides
L for mesh elements of 2178 and 6112. Again Fig. 2(b) illustrates
s¼p ð21Þ that mid-plane velocity profile for 2072 mesh elements almost
2
matches with the same for mesh elements 5992. Therefore,
The stream function can be given by,
mesh elements of 6112 and 5992 have been selected to carry
@w @w out the numerical simulation for triangular and semi-circular
U¼ ; V¼
: ð22Þ
@Y @X notched cavities respectively.
Pure mixed convection in a ferrofluid-filled lid-driven cavity 131

(a) (b)

Figure 2 Mid-plane velocity profile for different element numbers for (a) triangular and (b) semi-circular notched cavity at Gr = 104,
Ri = 0.1, Ha = 25 and u = 0.15.

Streamline Isotherm
Oztop et al. [43]
Present

Figure 3 Comparison of streamline and isotherms for Ra = 105, Re = 100 and Ha = 50 between numerical investigation by Oztop
et al. [46] and present study.

3.2. Code validation
The present code is validated with the previous investigation
by Oztop et al. [46] for lid-driven cavity with corner heater
under external magnetic field. Validation is established
through streamlines and isotherm contours. Comparison of
the flow and thermal field is illustrated in Fig. 3 for
Ra = 105, Re = 100, Ha = 50. From the figure it is significant
that the present code conforms to previous study and a good
comparison is registered. According to this, the present numer-
ical code and solution procedure are perfectly reliable. Hence,
numerical simulation can be carried out for the given study.
132
Ha = 10 Ha = 25
(a) (b)
(d) (e)
Figure 4 Effect of Hartmann number (Ha) on streamline for (a–c) triangular and (d–f) semi-circular notched cavity at Ri = 1,
u = 0 (—) and 0.15 (- - -).
4. Result and discussion
Primary objective of this numerical investigation is to com-
pare convective heat transfer scenario between a triangular
and a semi-circular notched cavity filled with Fe3O4–water
ferrofluid in the presence of external magnetic field.
Results from numerical simulation are presented in terms
of streamlines, isotherms and average Nusselt number dis-
tributions at the heated boundary for various values of
Richardson number (Ri), Hartmann number (Ha), and
nanoparticle volume fraction (u). Ri measures relative
importance of the natural and forced convection. By vary-
ing its numerical value (0.1 6 Ri 6 10), mode of convection
heat transfer is altered from forced to mixed to natural.
Ha represents strength of external magnetic field as body
force. To observe magnetic response of flow, Ha is varied
from 0 to 100.
Particular emphasis is put on observing pure mixed convec-
tion regime of heat transfer (Ri = 1). To compare effectiveness
of ferrofluid solid volume fraction of u = 0 (pure water) and
of u = 0.15 (ferrofluid) is selected for comparison of heat
transfer scenario.
K.Md. Rabbi et al.
Ha = 50
(c)
(f)
4.1. Effect of Hartmann number
Effect of Hartmann number on flow and thermal fields, for
both ferrofluid and pure water, is depicted in terms of stream
function and isothermal contours in Figs. 4 and 5 for triangu-
lar and semi-circular notched cavities respectively. Pure mixed
convective heat transport is selected as representative case. For
Ha = 10, a clockwise vortex is observed in the cavity and a
secondary vortex is observed near the left bottom end with
opposite sense of rotation. The pattern is similar for both types
of cavity and independent of fluid type. Increment of Ha to 25
generates two more secondary vortices around notch surface
due to the intense effect of magnetic field. However, the inten-
sity of these cells is negligible when compared with the main
circulation cells. Physically, high strength of magnetic field
tries to counter-balance the impact of forced convection. Thus
at high Ha, natural convection becomes marginally more dom-
inant compared to forced flow. Secondary vortices are the
results of secondary convection (natural mostly) inside the cav-
ity. When Ha = 50, the streamlines are stretched more, hori-
zontally, with lower strength and the secondary circulations
cover the notch surface. Streamlines also cluster around the
Pure mixed convection in a ferrofluid-filled lid-driven cavity
Ha = 10 Ha = 25
(a) (b)
(d) (e)
Figure 5 Effect of Hartmann number (Ha) on isothermal lines for (a–c) triangular and (d–f) semi-circular notched cavity at Ri = 1,
u = 0 (—) and 0.15 (- - -).
top surface. It is noticed that with the increase of Hartmann
number, value of stream function decreases. As increasing
Hartmann number induces stronger magnetic field, which in
turn produces Lorentz force in flow field, it decreases the
strength of flow (i.e. weak convection current) in the cavity.
The secondary circulation for ferrofluid (u = 0.15) stretches
more as it is more susceptible to magnetic field than pure water
(u = 0) in case of Ha = 25 and 50.
Stream function values in semi-circular notched cavity are
higher than the triangular one for high Hartmann numbers.
This denotes superiority of convection heat transfer in semi-
circular notched cavity than that in triangular one.
Impact of Hartmann number on thermal field is represented
by isotherm contours (see Fig. 5) for pure water (solid line) and
ferrofluid (dotted line) in triangular and semi-circular notched
cavities respectively.
For Ha = 10, approximately parallel isotherms are densely
distributed near the right half of heated notch due to the for-
mation of thermal boundary layer (i.e. weak convection).
Irregular, distorted and less dense isotherms evolve with the
approach from right to left half due to the increasing convec-
tive heat transfer. Higher Hartmann numbers, Ha = 25 and
50, show more parallel isotherms throughout the cavity. These
133
Ha = 50
(c)
(f)
horizontal isotherm contours indicate weaker convection and
domination of conduction heat transfer. It is observed that
with the increase of Ha, the isotherm contours become more
horizontal for ferrofluid compared to pure water as conduction
is dominant in ferrofluid than pure water.
As the Hartmann number increases, the effect of magnetic
field enhances. Thus, the more the Ha, the more the Lorentz
force dominates on the buoyancy force. Therefore, the strength
of flow and thermal field varies with the strength of magnetic
field.
4.2. Effect of Richardson number
Figs. 6 and 7 show evolution of streamline and isotherms with
Ri in triangular and semi-circular notched cavities respectively
at Ha = 25. To vary Ri, Grashof number is kept fixed while
changing Reynolds number through the lid velocity uo. Ri
value varies from 0.1 to 10 describing the transition between
forced and natural convection analysis.
For Ri = 0.1 one primary clockwise vortex and two sec-
ondary vortices on either side of the notch appear for triangu-
lar and semi-circular notched cavities. Forced convection,
134
Ri = 0.1 Ri = 1.0
(a) (b)
(d) (e)
Figure 6 Effect of Richardson number (Ri) on streamlines for (a–c) triangular and (d–f) semi-circular notched cavity at Ha = 25, u = 0
(—) and 0.15 (- - -).
which begets from inertia force due to driven velocity uo, is
dominant in this case. Further increase of Richardson number
(i.e. Ri = 1 and 10) gives augmented vortices. Both CW vor-
tices merge together and cover larger area of the cavity. With
the increase of Ri value, vortices absorb more energy from heat
source and become enlarged. Streamlines spread throughout
the cavity with higher stream function value due to higher
buoyancy force at high Ri. This indicates stronger convection
at higher Ri.
It is interesting to observe that for ferrofluid both the CW
and CCW vortices spread more compared to pure water.
Indeed faster heat transfer occurs for ferrofluid. It causes
enhancement of the buoyancy force and hence the flow inten-
sity. It is also noticed from analysis that the streamlines of
main cavity spread much in semi-circular notched cavity com-
pared to triangular one at higher strength. This indicates
higher convection effect in semi-circular than in triangular
notched cavity.
Fig. 7 shows that at Ri = 0.1, isotherms are almost uni-
formly distributed near the center of cavity which points to
dominance of conduction heat transfer. It is noticed that iso-
therms are bunched near the heat source and stretched toward
K.Md. Rabbi et al.
Ri = 10
(c)
(f)
right and left adiabatic walls. This pattern emerges due to
higher temperature gradient near the source and decreasing
temperature gradient away from source. For Ri = 1, the solid
isotherm contours (u = 0) become distorted. However, the
dotted isotherm contours (u = 0.15) are still almost uniform.
Further increase of Ri to 10 presents completely distorted iso-
therm patterns because of better convection. Isotherms seem to
follow the fluid flow path as we see that those are more densely
packed on left of heated notches.
Comparison of isotherms for ferrofluid and pure water
reveals that the temperature of ferrofluid is lower than that
of pure water at the vicinity of heat source and in right half
of cavity, but in left half of cavity lower temperature is
obtained for pure fluid. This is due to the presence of nanopar-
ticles of ferromagnetic materials enhances interaction among
recirculation zones.
Typically, the natural convection is negligible when
Ri < 0.1, forced convection is negligible when Ri > 10, and
neither is negligible when 0.1 < Ri < 10. Therefore,
0.1 < Ri < 10 is used to describe the effect of both natural
convection and forced convection on the flow and thermal
field.
Pure mixed convection in a ferrofluid-filled lid-driven cavity
Ri = 0.1 Ri = 1.0
(a) (b)
(d) (e)
Figure 7 Effect of Richardson number (Ri) on isothermal lines for (a–c) triangular and (d–f) semi-circular notched cavity at Ha = 25,
u = 0 (—) and 0.15 (- - -).
4.3. Impact of Ri on mid-plane velocity
Fig. 8 depicts effect of Ri on mid-plane x-velocity at different
Ha (10, 25 and 50) for triangular ((a), (c), and (e)) and semi-
circular ((b), (d), and (f)) notched cavities at Gr = 104 and
u = 0.15. As we see, increment of Ha, lessens the x-velocity
in lower half of the cavity while raising it on upper half. Same
pattern is observed for both kinds of notched cavities. Flow
behavior is entirely responsible for such velocity distribution.
Moreover, low Ri gives rise to overall high velocity of flow
inside the cavity.
4.4. Combined effect of Ha, Re and Ri on average Nusselt
number for ferrofluid (u = 0.15)
Average Nusselt number (Nu), as a function Ha and Re for fer-
rofluid (u = 0.15) at Gr = 103 and Gr = 104, is analyzed in
Figs. 9 and 10 for triangular ((a), (c)) and semi-circular ((b),
(d)) notched cavities respectively.
We see from Fig. 9(a)–(d) Nu decreases with the increment
of Ha. It conforms to the previous observation where it was
135
Ri = 10
(c)
(f)
found that Ha had an adverse effect on the flow strength inside
the cavity. Nu decreases rapidly up to Ha = 20. At higher
value of Hartmann number (Ha P 40), value of Nu becomes
almost constant. At higher Reynolds number (Re  500), ini-
tial average Nusselt number is high (Nu  10) which decreases
sharply up to Ha = 20. But at low Reynolds number
(Re  100), initial average Nusselt number is low (Nu  5)
which further decreases with Ha. Interestingly pattern of vari-
ation is quite similar for both types of cavities and both values
of Gr. Since taking only two values of Gr cannot be deemed as
conclusive, we can only comment on insensitivity of geometry
of notch on flow physics but not for increment of heat transfer
rate. This remark is backed by streamline plots previously
shown.
To observe impact of Ri further analysis is done by varying
its value at different Ha at Gr = 104 and u = 0.15 (see Fig. 10
(a) and (b)). It is observed that increasing Ri actually aids heat
transfer. Therefore, in case of convection in notched cavity, lid
velocity should be such that entire heat transfer process
remains more or less natural convective. Interestingly, forced
convection, which generally is supposed to augment heat trans-
fer rate, proves to be futile for the present case.
136 K.Md. Rabbi et al.

Ri = 0.1

(a) (b)
Ri = 1

(c) (d)
Ri = 10

(e) (f)

Figure 8 Effect of Richardson number (Ri) on mid-plane x-velocity for (a, c, e) triangular and (b, d, f) semi-circular notched cavity at
Gr = 104 and u = 0.15 .

4.5. Effect of solid volume fraction (u) on average Nusselt
number
Fig. 11 depicts the variation of average Nusselt number with
the increment of solid volume fraction of ferrofluid. It is signif-
icant that Nu increases with u no matter what strength the
external magnetic field possess. In case of triangular notched
cavity, Fig. 11(a) represents that better heat transfer rate (more
than 14%) is obtained (Nu  15) at lowest strength of mag-
netic field (Ha = 10). Moreover, Fig. 11(b) illustrates the same
for semi-circular shaped heater geometry. It is observed that
more than 30% of heat transfer rate can be achieved at
Ha = 10 with ferrofluid inside (u = 0.15). This phenomenon
is evident mainly because of the augmentation of thermal
Pure mixed convection in a ferrofluid-filled lid-driven cavity 137

(b)
(a)
Gr = 103

(d)
(c)
Gr = 104

Figure 9 Variation of average Nusselt number (Nu) with Hartmann number (Ha) and Reynolds number (Re) for (a, c) triangular and
(b, d) semi-circular notched cavity at u = 0.15.

(a) (b)

Figure 10 Variation of average Nusselt number (Nu) with Richardson number (Ri) for (a) triangular and (b) semi-circular notched
cavity at Gr = 104 and u = 0.15.
138
(a)
Figure 11 Variation of average Nusselt number (Nu) with solid volume fraction (u) for (a) triangular and (b) semi-circular notched
cavity at Gr = 104 and Ri = 1.
conductivity and energy transfer with the increment of solid
concentration. Increment of flow intensity is also obtained
with higher solid volume fraction. As the solid volume fraction
is increased, micro-convection becomes more prominent which
contributes to overall heat transfer rate.
Thus it can be concluded that implementation of ferrous
particles on water in an enclosure of any heater geometry sud-
denly enhances the heat transfer while analyzing pure mixed
convection (Ri = 1).
5. Conclusions
In present analysis, both geometric and physical behaviors of
mixed convection in a square notched cavity under magnetic
field are highlighted in brief manner. This study reveals some
key aspects of such convective flow. It is revealed that Ha
has negative impact on convection. This conclusion is absolute
and is not dependent on notch geometry or convection regime.
Ferrofluid is superior in terms of heat transfer augmentation to
pure water in all cases of analyzed scenario. Interestingly, in
notch induced convection, natural convection proved to be
more conducive mode of heat transfer compared to forced
convection. Impact of Ha is most severe among other factors.
So by tuning Ha as key parameter, heat transfer rate can be
controlled. Finally, it is concluded that semi-circular notched
cavity seems to produce higher value of Nusselt number com-
pared to triangular notched cavity. This is expected since,
semi-circular heater has more area for heat transfer for identi-
cal geometry. In fact, heat transfer increases up to 14% and
30% due to the implementation of ferrofluid for triangular
and semi-circular notched cavities respectively. Present analy-
sis can be taken ahead further by implementing experimental
measures to observe the actual scenario.
References
[1] C. Vives, C. Perry, Effects of magnetically damped convection
during the controlled solidification of metals and alloys, Int. J.
Heat Mass Trans. 30 (1987) 479–496.
K.Md. Rabbi et al.
(b)
[2] H.P. Utech, M.C. Flemmings, Elimination of solute banding in
indium antimonide crystals by growth in a magnetic field, J.
Appl. Phys. 37 (1966) 2021–2024.
[3] J.C. Umavathi, A.J. Chamkha, A. Mateen, J.P. Kumar,
Unsteady magnetohydrodynamic two fluid flow and heat
transfer in a horizontal channel, Heat Technol. 26 (2008) 133–
145.
[4] H. Saleh, R. Roslan, I. Hashim, Natural convection in a porous
trapezoidal enclosure with an inclined magnetic field, Comput.
Fluids 47 (2011) 155–164.
[5] P.X. Yu, J.X. Qiu, Q. Qin, Z.F. Tian, Numerical investigation of
natural convection in a rectangular cavity under different
directions of uniform magnetic field, Int. J. Heat Mass Trans.
67 (2013) 1131–1144.
[6] M.C. Ece, E. Buyuk, Natural-convection flow under a magnetic
field in an inclined rectangular enclosure heated and cooled on
adjacent walls, Fluid Dyn. Res. 38 (2006) 564–590.
[7] V.A.F. Costa, A.C.M. Sousa, P. Vasseur, Natural convection in
square enclosures filled with fluid-saturated porous media under
the influence of the magnetic field induced by two parallel
vertical electric currents, Int. J. Heat Mass Trans. 55 (2012)
7321–7329.
[8] N. Rudraiah, R.M. Barron, M. Venkatachalappa, C.K.
Subbaraya, Effect of magnetic field on free convection in a
rectangular enclosure, Int. J. Eng. Sci. 33 (1995)
1075–1084.
[9] H.F. Oztop, M. Oztop, Y. Varol, Numerical simulation of
magnetohydrodynamic buoyancy-induced flow in a non-
isothermally heated square enclosure, Commun. Nonlinear Sci.
Numer. Simul. 14 (2009) 770–778.
[10] H.F. Oztop, M.M. Rahman, A. Ahsan, M. Hasanuzzaman, R.
Saidur, K. Al-Salem, N.A. Rahim, MHD natural convection in
an enclosure from two semi-circular heaters on the bottom wall,
Int. J. Heat Mass Trans. 55 (2012) 1844–1854.
[11] M. Sheikholeslami, M. Gorji-Bandpy, D.D. Ganji, MHD free
convection in an eccentric semi-annulus filled with nanofluid, J.
Taiwan Inst. Chem. Eng. 45 (2014) 1204–1216.
[12] M. Sheikholeslami, M. Gorji-Bandpy, D.D. Ganji, Numerical
investigation of MHD effects on Al2O3–water nanofluid flow
and heat transfer in a semi-annulus enclosure using LBM,
Energy 60 (2013) 501–510.
[13] M.S. Hossain, M.A. Alim, MHD free convection within
trapezoidal cavity with non-uniformly heated bottom wall, Int.
J. Heat Mass Trans. 69 (2014) 327–336.
Pure mixed convection in a ferrofluid-filled lid-driven cavity
[14] S.P. Jang, S.U.S. Choi, Role of Brownian motion in the
enhanced thermal conductivity of nanofluids, Appl. Phys. Lett.
84 (2014) 4316–4318.
[15] J. Koo, C. Kleinstreuer, A new thermal conductivity model for
nanofluids, J. Nanoparticles Res. 6 (2004) 577–588.
[16] W. Evans, J. Fish, P. Keblinski, Role of Brownian motion
hydrodynamics on nanofluid thermal conductivity, Appl. Phys.
Lett. 88 (2006).
[17] R. Prasher, P. Bhattacharya, P.E. Phelan, Brownian-motion-
based convective-conductive model for effective thermal
conductivity of nanofluids, J. Heat Trans. 128 (2006) 588–595.
[18] M.M. Rahman, S. Mojumder, S. Saha, S. Mekhilef, R. Saidur,
Augmentation of natural convection heat transfer in triangular
shape solar collector by utilizing water based nanofluids having
a corrugated bottom wall, Int. Commun. Heat Mass Trans. 50
(2015) 117–127.
[19] A. Gupta, R. Kumar, Role of Brownian motion on the thermal
conductivity enhancement of nanofluids, Appl. Phys. Lett. 91
(2006).
[20] N.S. Bondareva, M.A. Sheremet, Magnetic field effect on the
unsteady natural convection in a right-angle trapezoidal cavity
filled with a nanofluid: Buongiorno’s mathematical model, Int.
J. Numer. Met. Heat Fluid Flow 25 (2015) 1924–1946.
[21] A.H. Mahmoudia, I. Pop, M. Shahi, F. Talebi, MHD natural
convection and entropy generation in a trapezoidal enclosure
using Cu–water nanofluid, Comput. Fluids 72 (2013) 46–62.
[22] M. Mohammadpourfard, Numerical study of ferrofluid flow
and heat transfer in the presence of a non-uniform magnetic field
in rectangular microchannels, Heat Trans. Asian Res. 41 (2012)
302–317.
[23] M. Sheikholeslami, D.D. Ganji, Ferrohydrodynamic and
magnetohydrodynamic effects on ferrofluid flow and
convective heat transfer, Energy 75 (2014) 400–410.
[24] M. Sheikholeslami, M. Gorji-Bandpy, Free convection of
ferrofluid in a cavity heated from below in the presence of an
external magnetic field, Powder. Technol. 256 (2014) 490–498.
[25] G. Kefayati, Natural convection of ferrofluid in a linearly heated
cavity utilizing lbm, J. Mol. Liq. 191 (2014) 1–9.
[26] M. Ashouri, B. Ebrahimi, M. Shafii, M.H. Saidi, M.S. Saidi,
Correlation for Nusselt number in pure magnetic convection
ferrofluid flow in a square cavity by a numerical investigation, J.
Magn. Magn. Mater. 322 (2010) 3607–3613.
[27] A. Gavili, M. Lajvardi, J. Sabbaghzadeh, The effect of magnetic
field gradient on ferrofluids heat transfer in a two-dimensional
enclosure, J. Comput. Theor. Nanosci. 7 (2010) 1425–1435.
[28] M. Lajvardi, J. Moghimi-Rad, I. Hadi, A. Gavili, T.D. Isfahani,
F. Zabihi, J. Sabbaghzadeh, Experimental investigation for
enhanced ferrofluid heat transfer under magnetic field effect, J.
Magn. Magn. Mater. 322 (2010) 3508–3513.
[29] N.S. Bondareva, M.A. Sheremet, Influence of uniform magnetic
field on laminar regimes of natural convection in an enclosure,
Thermophy. Aeromech. 22 (2015) 203–216.
[30] T.K. Aldoss, Y.D. Ali, M.A. Al-Nim, MHD mixed convection
from a horizontal circular cylinder, Numer. Heat Trans. 30
(1996) 379–396.
139
[31] A. Kaya, The effect of conjugate heat transfer on MHD mixed
convection about a vertical slender hollow cylinder, Commun.
Nonlinear Sci. Numer. Simul. 16 (2011) 1905–1916.
[32] A.J. Chamkha, A.M. Aly, H.F. Al-Mudhaf, Laminar MHD
mixed convection flow of a nanofluid along a stretching
permeable surface in the presence of heat generation or
absorption effects, Int. J. Micro-Nano Thermal Fluid Trans.
Phenom. 2 (2011) 51–70.
[33] M.M. Rahman, H.F. Oztop, R. Saidur, S. Mekhilef, K. Al-
salem, Finite element solution of MHD mixed convection in a
channel with a fully or partially heated cavity, Comput. Fluids
79 (2013) 53–64.
[34] G.H.R. Kefayati, M. Gorjy-Bandpy, H. Sajjadi, D.D. Ganji,
Lattice Boltzman simulation of MHD convection in a lid-driven
cavity with linearly heated wall, Sci. Iranica 19 (2012) 1053–
1065.
[35] K. Al-Salem, H.F. Oztop, I. Pop, Y. Varol, Effects of moving lid
direction on MHD mixed convection in a linearly heated cavity,
Int. J. Heat Mass Trans. 55 (2012) 1103–1112.
[36] J.C. Umavathi, M.S. Malashetty, Magnetohydrodynamic mixed
convection in a vertical channel, Int. J. Non-Linear Mech. 40
(2005) 91–101.
[37] E.A. Nada, A.J. Chamkha, Mixed convection flow of a
nanofluid in a lid-driven cavity filled with a wavy wall, Int.
Commun. Heat Mass Trans. 57 (2014) 36–47.
[38] B. Ghasemi, S.M. Aminossadati, Mixed convection in a lid-
driven triangular enclosure filled with nanofluids, Int. Commun.
Heat Mass Trans. 37 (2010) 1142–1148.
[39] H.M. Elshehabey, S.E. Ahmed, MHD mixed convection in lid-
driven cavity filled by a nanofluid with sinusoidal temperature
distribution on the both vertical walls using Buongiorno’s
nanofluid model, Int. J. Heat Mass Trans. 88 (2015) 181–202.
[40] M.A. Sheremet, I. Pop, Mixed convection in a lid-driven square
cavity filled by a nanofluid: Buongiorno’s mathematical model,
Appl. Math. Comput. 266 (2015) 792–808.
[41] H. Aminfar, M. Mohammadpourfard, S.A. Zonouz, Numerical
study of the ferrofluid flow and heat transfer through a
rectangular duct in the presence of a non-uniform transverse
magnetic field, J. Magn. Magn. Mater. 327 (2013) 31–42.
[42] H. Aminfar, M. Mohammadpourfard, F. Mohseni, Two-phase
mixture model simulation of the hydro-thermal behavior of an
electrical conductive ferrofluid in the presence of magnetic fields,
J. Magn. Magn. Mater. 324 (2012) 830–842.
[43] B. Ghasemi, S.M. Aminossadati, A. Raisi, Magnetic field effect
on natural convection in a nanofluid-filled square enclosure, Int.
J. Therm. Sci. 50 (2011) 1748–1756.
[44] H.C. Brinkman, The viscosity of concentrated suspensions and
solution, J. Chem. Phys. 20 (1952) 571–581.
[45] J. Maxwell-Garnett, Colours in metal glasses and in metallic
films, Phil. Trans. Roy. Soc. Lond. A 203 (1904) 385–420.
[46] H.F. Oztop, K. Al-Salem, I. Pop, MHD mixed convection in lid-
driven cavity with corner heater, Int. J. Heat Mass Trans. 54
(2011) 3494–3504.