International Journal of Thermal Sciences 50 (2011) 1984e1995

Contents lists available at ScienceDirect

International Journal of Thermal Sciences

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

Non-Newtonian fluid flow in plane channels: Heat transfer

enhancement using porous blocks
R. Nebbali*, K. Bouhadef
Laboratoire des Transports Polyphasiques et Milieux Poreux, FGMGP-USTHB, BP 32 El-Alia 16111 Algiers, Algeria

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

Article history: A numerical investigation is performed for heat transfer enhancement in a parallel-plate channel. Porous
Received 30 June 2010 blocks are inserted to partially fill the channel, which is crossed by a power-law fluid. The modified
Received in revised form BrinkmaneForchheimer extended Darcy model for power-law fluids is used in the porous layer while the
12 April 2011
NaviereStokes equation are employed in the clear region of the channel. Results are reported for two
Accepted 13 April 2011
cases: (1) a channel with a single porous block and (2) a channel with two porous blocks mounted
Available online 28 May 2011
alternatively at the bottom and top walls of the channel. The combined effects of both the porous blocks
and the non-Newtonian fluid properties on the hydrodynamic and thermal characteristics of the flow are
Keywords:
Non-Newtonian fluid
analyzed. To this end, computations are performed to highlight the effect of parameters such as the Darcy
Porous blocks number, the Reynolds number, and the power-law index. For the case of the single porous block, the heat
Forced convection transfer is enhanced and maximized at low permeability. In the second configuration as well, heat
Heat transfer enhancement transfer improvement can also be obtained, even if the thermal conductivity ratio is equal to 1. This
enhancement is maximized for particular values of the Darcy number that depend on the flow pattern
and the power-law index. Concerning the rheological aspect, it is found that pseudo plastic fluids exhibit
the highest Nusselt number and the lowest pressure loss. The presence of the porous inserts causes
a significant increase in pressure drop, which is found to be more important with dilatant fluids.
Ó 2011 Elsevier Masson SAS. All rights reserved.

1. Introduction and Al-Nimr [2] numerically explored heat transfer enhancement

During the last few years, there has been an increasing interest
in fundamental studies of forced convection in ducts of various
shapes fully or partially filled with a porous material. This interest is
due to the presence of porous media in numerous engineering
applications and natural processes such as ceramic processing,
filtration, geothermal systems, groundwater flow, enhanced oil
recovery, compact heat exchangers, packed bed chemical reactors
and many others. The use of porous/fluid composite systems is an
innovative method able to provide valuable solutions for improving
energy the efficiency of thermal systems. This can have a positive
impact in areas ranging from preservation of energy resources to
limiting global warming. Hadim [1] conducted a numerical analysis
on laminar forced convection in fully and partially porous parallel-
plate channels with discrete heat sources. The obtained results
showed that the partially porous channel can get almost the same
improvement of heat transfer as that obtained with the fully porous
channel while the pressure drop is drastically reduced. Hamdan
* Corresponding author.
E-mail addresses: rnebbali@usthb.dz (R. Nebbali), kbouhadef@usthb.dz (K. Bouhadef).
by using a succession of high-thermal conductivity porous fins in
isothermal parallel-plate channel. It was found that the heat
transfer can be enhanced by using high thermal conductivity fins
with low permeability and by increasing the microscopic inertial
coefficient. Using a vorticity stream function formulation, Huang
and Vafai [3] showed that a significant heat transfer increase can be
obtained by adding porous blocks on the bottom wall of an
isothermal parallel-plate channel. Sung et al. [4] reported some
results on the effect of both height and permeability of a single
porous block on the flow and heat transfer characteristics of forced
convection in a channel. They considered two types of location for
the porous block. The situation where the porous medium is
inserted between the channel walls has been analyzed by Tong
et al. [5]. It was shown that heat transfer enhancement can be
obtained. Moreover, either the partially or the completely porous
channel according to the choice of the physical conditions specific
to each configuration, allows obtaining a maximum heat transfer.
Chikh et al. [6] focused their attention on an annular cylindrical
duct with a porous substrate mounted on the inner cylinder. The
velocity field was obtained analytically for a fully developed flow.
The results showed that there exists a critical thickness of the
porous layer at which heat transfer is minimal. Nield and Kuznetsov

1290-0729/$ e see front matter Ó 2011 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ijthermalsci.2011.04.013
 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995 1985

Nomenclature S Dimensionless block height

u Axial velocity
Cp Specific heat U Dimensionless axial velocity
CF Inertial factor Uo Inlet velocity
C* Modified inertial factor v Transverse velocity
Ct Turtuosity factor V Dimensionless transverse velocity
d Spacing between the blocks x Axial coordinate
Da Modified Darcy number, ðK * Þ2=nþ1 =H 2 X dimensionless axial coordinate.
Dh Hydraulic diameter, 2H y Transverse coordinate
f Fanning factor, ðdp=dxÞDh =2 rU02 Y Dimensionless transverse coordinate
h Convection heat transfer coefficient wb Block width
H Channel height Wi Dimensionless width of the i th block
J Second invariant of the strain-rate tensor
ke Effective thermal conductivity Greek symbols
kf Fluid thermal conductivity a Variable defined in (15)
K Permeability 3 Porosity
K* Modified permeability m* Consistency factor
l1 Entrance length q Dimensionless temperature
l2 Length after last block qb Dimensionless bulk temperature
n Power-law index qw Dimensionless wall temperature
N Number of block F Dimensionless second invariant of the strain-rate
n0 Modified power-law index tensor
Nui Average Nusselt number on a block l Binary variable
Num Average Nusselt number r Fluid density
Nux Local Nusselt number J Dimensionless stream function
p Pressure
P Dimensionless pressure Subscripts
qw Heat flux density b Bulk
Pr Generalized Prandtl number, ðm* Cp =kf ÞðU0 =HÞn1 e Effective
Re Modified Reynolds number, rU02n H n =m* f Fluid
Rk Thermal conductivity ratio, ke =kf m Average
T Temperature o Inlet
To Inlet temperature w Wall
s Block height

[7] analyzed analytically the flow in a multi layered parallel-plate
porous channel, which consists of two adjacent porous layers fol-
lowed by a clear region. The permeability is constant in the porous
layer attached to the top wall while in the second porous layer the
permeability varies linearly across the channel. The obtained
boundary friction factors agree with those given by a model that
uses the Beavers-Joseph condition at the interface between a Darcy
porous medium and a clear fluid. Ould-Amer et al. [8] analyzed the
effect of a porous insert between heated solid blocks in a parallel-
plate channel. They showed in particular that the porous insert
enhances the heat transfer rate on the vertical sides of the blocks.
Chikh et al. [9] reported a numerical study of forced convection
enhancement in a parallel-plate channel intermittently heated by
using porous blocks mounted on the heated parts of the channel.
They showed that the blocks may alter substantially the flow
pattern and improve the heat transfer under appropriate condi-
tions. The thermal conductivity ratio is another important param-
eter that has a positive effect on heat transfer as shown in Alkam
et al. [10]. This work focuses on the transient forced convection in
a partially porous parallel-plate channel. It highlights that the heat
exchange can be improved by adopting a high thermal conductivity
ratio. The use of porous baffles to enhance heat transfer was dis-
cussed by Miranda and Anand [11]. The authors have inserted
sixteen porous baffles mounted alternatively on the upper and the
lower walls of a channel; they analyzed their impact on both the
dynamic and the thermal behaviors of the flow. The results
revealed that the porous baffles outperform solid baffles from
a pressure drop point of view, while they underperform solid
baffles from the heat transfer aspect. Huang et al. [12] attached
porous covers to the heated blocks in a parallel-plate channel to
dissipate more efficiently the heat generated by the blocks. It was
shown that the addition of the porous covers enhance significantly
the cooling of the heated blocks. Recently, Li et al. [13] numerically
studied the convection heat transfer in a channel with staggered
porous blocks. The effects of Darcy number, Reynolds number,
porous block height, and width on the fluid flow and heat transfer
characteristics were analyzed. It was shown that the heat transfer is
significantly enhanced with the decrease of Darcy number at the
expense of the high pressure drop. It was also revealed that the
height and width of the porous blocks can be optimized for high
heat transfer performance. Huang et al. [14] focused their attention
on the concept of heat transfer enhancement in the core flow of
a channel. They considered a tube under uniform heat flux condi-
tion in which porous inserts with a slightly smaller diameter were
placed in the core flow. Both numerical and experimental results
were reported and it was shown that the convective heat transfer is
considerably enhanced by the porous inserts and the correspond-
ing flow resistance increases in a reasonable extent. Huang and
Yang [15] presented a numerical investigation of the flow field and
heat transfer characteristics of two successive porous-block-
mounted heat sources subjected to pulsating channel flow. It was
shown that specific choices in some governing parameters can
significantly enhance the cooling of the heat sources. Satyamurty
and Bhargavi [16] considered a parallel-plate channel with a porous
1986 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995

medium attached to the lower wall. While the flow was considered simultaneous presence of the non-Newtonian fluid and the porous
developed, the thermal establishment was analyzed. It was shown matrices on both fluid flow and heat transfer characteristics. To this
that maximum heat transfer enhancement is attained for a porous end, the influence of various parameters governing the hydrody-
thickness equal to 80% of the channel height. Moreover, the effect of namic and thermal fields such as the Reynolds number, the Darcy
the stress jump at the interface on both optimum porous fraction number and the power-law index, are examined to establish the
and enhancement in heat transfer was found to be not significant. fundamental effects and provide practical results.
The above literature review reveals that most of the undertaken
works were devoted to Newtonian fluids. However, in numerous
2. Mathematical formulation
engineering applications, the fluids often drift away from the
Newtonian behavior. Thus, a growing number of authors take into
As stressed earlier, changes in flow structure and its influence on
account the rheological aspect by considering non-Newtonian
heat transfer is an important aspect that must be analyzed deeply.
fluids. Chen and Hadim [17] analyzed numerically the forced
The geometrical configuration under investigation in this work,
convective flow of a power-law fluid in a channel completely filled
including the heating zone, and the porous blocks location, are
with a porous medium. It was demonstrated that the decrease of
depicted in Fig. 1. The flow is considered steady, incompressible,
the particle diameter or the power-law index increases the velocity
laminar and two dimensional. The inlet fluid temperature and
and temperature gradients, resulting in an increase in the Nusselt
velocity are assumed to be uniform. The thermophysical properties
number. The presence of the porous medium magnifies the pres-
of the fluid and the porous inserts are held constant except for the
sure drop which is significantly greater for dilatants fluids. The
variation of the viscosity of the power-law fluid which depends on
transient forced convection in a porous annular channel in the
shear rate. The porous medium is assumed homogeneous, isotropic,
presence of a power-law fluid was investigated by Alkam et al. [18].
and in local thermodynamic equilibrium with the fluid. The flow in
It was shown that increasing permeability or decreasing the index n
the porous medium is modeled using a modified Brinkmane
decreases the Nusselt number and the establishment length,
Forchheimer-extended Darcy model for power-law fluids, which
whereas the opposite effect is observed when the Peclet number
takes into account boundary and inertia effects as reported by
increases. Al-Nimr and Aldoss [19] investigated the impact of
Shenoy [23].
inertial effects on the flow of a non-Newtonian fluid in a channel
In the clear region, the flow field is governed by the Naviere
completely filled by a porous medium. It was demonstrated that
Stokes equations while the thermal field is governed by the energy
above a particular value, the effect of the power-law index on the
equation. Tacking into account the previously adopted hypothesis,
Forchheimer term can be neglected. Nebbali and Bouhadef [20]
the governing equations are:
performed a numerical investigation of a fully developped forced
Continuity:
convection in a 3D square duct partially filled with a porous
substrate in the presence of a power-law fluid. They established vu vv
that there are particular values of the thickness and the perme- þ ¼ 0 (1)
vx vy
ability beyond which the heat transfer is improved. Furthermore, it
was shown that dilatant fluids allow obtaining the best thermal x momentum equation:
performance while, from a dynamic point of view, these fluids

 
 
vu vu vp v vu v vu vv
exhibit the highest pressure drop. Recently, Aksoy and Pakdemirli r u þv ¼ þ m* 2J þ m* J þ (2)
[21] reported an analysis of a non-Newtonian fluid flow through vx vy vx vx vx vy vy vx
a porous parallel-plate channel. They considered a third-grade fluid y momentum equation:
and the channel walls were kept at different temperatures.


 
 
Approximate analytical solutions were obtained for the velocity vv vv vp v vv vu v vv
profile and the temperature distribution by using perturbation r u þv ¼ þ m* J þ þ m* 2J (3)
vx vy vy vx vx vy vy vy
techniques. It was shown that the used approach shows good
agreement with numerical solutions. Chen et al. [22] used a gray Energy equation:
lattice Boltzmann model for the simulation of a power-law fluid   !
through a parallel-plate channel partly filled with a porous vT vT v2 T v2 T
rCp u þv ¼ kf þ (4)
medium. It was found that the slip velocity of the power-law fluid vx vy vx2 vy2
at the porous interface increases with the power index for a given
porous medium, and increases with the porosity of the porous According to [23], treating the fluid saturated porous medium
media for a given power-law fluid. as a continuum, the local-volume averages of the conservation
In most of the studies previously mentioned, the situation equations for mass, momentum, and energy in the porous region
where the porous blocks are mounted alternatively on the bottom are:
and the top walls of a parallel-plate channel is not well docu-
mented. Such configurations disrupt significantly the structure of Heat Source Porous medium
the flow and, so can significantly influence the heat transfer. The
a
Uo, To
present paper presents a numerical investigation of fluid flow and s H
forced convection heat transfer of a power-law non-Newtonian y, x,
fluid (milk, gelatin, blood and liquid cement, suspension of rice l1 l2
wb
starch) in a parallel-plate channel partially obstructed with porous
blocks. Two interior shape types are considered for the channel: b
a single porous block that occupies the whole height of the channel, Uo, To s
H
referred to as type (A), and two porous blocks mounted alterna- y, x, s
tively on the top and the bottom walls of the channel, referred as
type (B). Local heat sources are placed at the bottom wall of l1 wb d wb l2
the channel in the location of the porous blocks. The purpose of the
present study is to explore the combined effects induced by the Fig. 1. Flow configurations, (a) single porous block, (b) two porous blocks.
 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995 1987

Continuity: Inlet:

vu vv uð0; yÞ ¼ U0 ; vð0; yÞ ¼ 0; Tð0; yÞ ¼ T0 (13)

þ ¼ 0 (5)
vx vy
Top wall:
x momentum equation:
vT

 uðx; HÞ ¼ vðx; HÞ ¼ 0; ¼ 0 (14)
r vu vu ffin1
vp m* pffiffiffiffiffiffiffiffiffiffiffiffiffiffi rC pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi vy y¼H
u þv ¼  * u2 þ v2 u  pffiffiffiFffi u2 þ v2 u
3 vx vy
2 vx K K

 
  Bottom wall:
m* v vu m* v vu vv
þ n 2J þ n J þ (6)
3 vx vx 3 vy vy vx vT
uðx; 0Þ ¼ vðx; 0Þ ¼ 0; kf ¼ aqw (15)
vy y¼0
y momentum equation:

 Outlet:
r vv vv ffin1
vp m* pffiffiffiffiffiffiffiffiffiffiffiffiffiffi rC pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u þv ¼  * u2 þ v2 v  pffiffiffiFffi u2 þ v2 v vu vT
32 vx vy vy K K ¼ 0; v ¼ 0; ¼ 0 (16)

 
  vx vx
m* v vv m* v vv vu
þ n 2J þ n J þ (7) The variable a is equal to 1 for axial positions located in the
3 vy vy 3 vx vx vy
heated regions. Elsewhere, this variable is set to zero.
Energy equation: The governing equations are cast in dimensionless form by
  ! adopting the following non-dimensional variables:
vT vT v2 T v2 T
rCp u þv ¼ ke þ (8)
vx vy vx2 vy2 x y u v p T To
X ¼ ; Y ¼ ; U ¼ ; V ¼ ; P ¼ . ; q ¼ (17)
H H U0 U0 rf U02 32 ðqw HÞ=kf
In the momentum equations, K is the intrinsic permeability and
K* is the modified permeability for non Newtonian fluids which The introduction of a binary variable l allows the use of a single
depends on the structure of the porous medium and on the power- set of governing equations for both fluid and porous region. With
law index of the fluid and is given by Christopher and middleman this formulation, the interface matching conditions are satisfacto-
[24]: rily dealt as in Patankar [26]; the resulting numerical solution
  procedure is greatly simplified. Several authors [1,4,27] have
1  n3 n 50K ðnþ1Þ=2 successfully used this approach. Consequently, the dimensionless
K* ¼ (9)
2Ct 3n þ 1 33 equations are written as follows:
where 3 is the porosity of the porous medium and Ct is the tur- Continuity:
tuosity factor which is defined by several expressions in the liter-
vU vV
ature. In the present study, the expression given by Dharmadhikari þ ¼ 0 (18)
vX vY
and Kale [25] is adopted:
X- momentum equation:
 n0   3ð10n0 3Þ=ð10n0 þ11Þ
2 8n0 10n0  3 75 2 3
Ct ¼ (10)
3 9n0 þ 3 6n0 þ 1 16     vP ! n1 !
vU vU 6 jV j C j V j7
*
U þV ¼  l 1  32 þ 32  l32 6
4 þ pffiffiffiffiffiffi 7 U
When a power-law fluid flows through a porous medium, it vX vY vX Da 5
n1
ReDa 2
tends to follow a path which is either shorter or longer due to its
rheological character. To account for this effect, a modified     v  F
vU vV 
expression of the power-law index is used in the expression of the þ l 32n  1 þ 1 þ
vY Re vY vX
turtuosity factor. As reported in [17], it is given, by:     v  F vU 
þ l 32n  1 þ 1 2 ð19Þ
n0 ¼ n þ 0:3ð1  nÞ (11) vX Re vX

J is the second invariant of the strain-rate tensor and is given Y- momentum equation:
by: 2 3
   ! n1
 vP !
vV vV 6 jV j C j V j7
*
(   n1 U þV ¼  l 1 32 þ 32  l32 6 þ pffiffiffiffiffiffi 7
 2   ) vX vY vY 4 n1
Da 5
V
vu 2 vv vu vv 2 2 ReDa
2

J ¼ 2 þ2 þ þ (12)
vx vy vy vx     v  F
vU vV 
þ l 32n 1 þ1 þ
The boundary conditions are such that a non-slip condition vX Re vY vX
    v  
occurs at the impermeable walls which are adiabatic except at the F vV
porous block location where a uniform heat flux is specified. At the
þ l 32n 1 þ1 2 ð20Þ
vY Re vY
channel inlet, the fluid enters the domain with a uniform velocity
and a uniform temperature. At fluideporous interfaces, matching Energy equation:
conditions are prescribed to ensure the continuity of velocity !
components, pressure, stress, temperature and heat flux. The vq vq lðRk  1Þ þ 1 v2 q v2 q
downstream length of the channel is taken sufficiently long to U þV ¼ þ (21)
vX vY RePr vX 2 vY 2
ensure that fully developed conditions are applicable at the exit.
Then, the relevant boundary conditions which complement the the parameter l is set equal to 1 in the porous region and 0 in the
model are summarized as follows: fluid region.
1988 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995

In the above equations, F is the dimensionless form of the 1,0

second invariant of the strain-rate tensor while C* the modified
inertial coefficient defined, by:

Transversal position, Y
0,8
 1=ðnþ1Þ Equation (33)
* * 1=2 n=0.5
C ¼ CF K =K
n=1.5
0,6
The dimensionless boundary conditions are as follows: n=1
Present work
Inlet: n=1
n=1.5
0,4
Uð0; YÞ ¼ 1; Vð0; YÞ ¼ 0; qð0; YÞ ¼ 0 (22) n=0.5

Top wall:
0,2

vq
UðX; 1Þ ¼ 0; VðX; 1Þ ¼ 0; ¼ 0 (23)
vY Y¼1 0,0
0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8
Bottom wall:
Axial velocity, U
vq
UðX; 1Þ ¼ 0; VðX; 1Þ ¼ 0; ¼ a (24)
vY Y¼0 Fig. 2. Developed velocity profile for different n, Da / N and Re ¼ 100, comparison
with analytical solution.
Outlet:

vU vq
¼ 0; V ¼ 0; ¼ 0 (25) 3. Numerical solution procedure
vX vX
The pressure drop increases as a result of inserting the porous The governing equations were solved numerically by using the
blocks in the channel. This can be regarded as the price to pay for SIMPLE algorithm to link the pressure and velocity fields as out-
the gain in heat transfer enhancement due to the porous inserts. lined in [26]. It is based on the finite volume formulation, which
This pressure drop is calculated in terms of the Fanning factor f, ensures conservation of mass, momentum and energy over each
which is defined as follows:

dp
Dh
f ¼  dx 2 (26)
2rU0

Thereafter, to assess the effect of the porous substrate on the

heat transfer, the local Nusselt number along the surface of the
blocks is defined as:

hH
Nux ¼ (27)
kf

while the average Nusselt number on each porous block is calcu-

lated as follow:
Z
1
Nui ¼ Nux dx (28)
Wi
and, the global average Nusselt number is given by:

1X
Num ¼ Nui (29)
N
Using the dimensionless variables, the last two parameters are
written:

dP
f ¼ 32 (30)
dX

vq
½ðRk  1Þl þ 1
vY Y¼0
Nux ¼  (31)
qw  qb
where Rk is the thermal conductivity ratio, while qb is the average
bulk temperature given by the dimensionless expression:
Z
U qdY
qb ¼ Z (32)
UdY Fig. 3. Comparison of the streamlines contours with that obtained by Sung et al. [4]
for (a) Re ¼ 100, Da ¼ 105, S ¼ 0.9 and (b) Re ¼ 100, Da ¼ 105, S ¼ 0.5.
 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995 1989

control volume as well as across the fluid/porous interface as 10

7

well. The power law scheme was used to model the interaction
n=1

Average friction coefficient, fRem

6
between the convection and diffusion terms. The nonlinear terms 10
n=1.5
in the momentum equations due to the non-Newtonian fluid and
5
n=0.5
the presence of the porous material were treated as source terms 10
and linearized as described in [26]. The mesh required in order to
4
obtain acceptable accuracy was found by trial and error. It was 10
established that a 60* 220 grid yielded results with less than 2%
3
discrepancy with those obtained by using an 80*320 grid. The 10
accuracy of the results with a 120*440 grid is improved by 0.6%
2
with that obtained by the 80*320 grid. Due to the large variation 10
of the physical quantities in certain regions of the channel,
1
a skewed grid distribution is used along both coordinate axes. In 10
the X-direction, the mesh points are densely packed at the
0
channel entrance. In the Y-direction, grid points are clustered 10 -6 -5 -4 -3 -2 -1 0
10 10 10 10 10 10 10
near the top and bottom surfaces of the channel as well as in the
neighborhood of the fluid/porous block interfaces in both Darcy number, Da
directions.
Fig. 5. Effects of the index n on the variations of the average friction coefficient with
The accuracy and the validity of the numerical developed model
Darcy number for Re ¼ 100.
were checked by comparing the present results with relevant
limiting cases available in the literature. A first comparison was
performed on the developed velocity profile at the channel exit for
2 3
the non porous case and for different values of the power-law nþ1  nþ1
2n þ 1 nþ1 1
n n
index. To this end, a large value was assigned to the permeability n 6 1 7
UðYÞ ¼  2 4 Y þ  5 (33)
while the porosity was set to 1. The developed velocity profile is nþ1 2 2
obtained by solving analytically the simplified X-momentum
equation, and is given by: As shown in Fig. 2, the numerical velocity profiles obtained for
several values of n agree very well with those given by the
analytical solution (33).
Second, the flow pattern was compared with the results of Sung
a 0,20 et al. [4] for a single porous block which partially obstructs the
Present work channel height. As it can be seen in Fig. 3, the obtained streamlines
Lower wall temperature, θ P

Re=100 are similar to those reported in [4] for the Newtonian case.
0,15 Re=250
Re=500
Last, Fig. 4 shows a comparison of the temperature distribution
Sung et al. [4] at the channel wall with those outlined in [4]. A very good agree-
Re=100
Re=250 ment is observed for different values of the Reynolds number at
0,10
Re=500 both walls.

0,05
4. Results and discussion

0,00 This section presents results obtained from the numerical

simulations, which highlight the effects of the governing parame-
ters, such as Darcy number, Reynolds number, power-law index, on
-0,05
2,5 5,0 7,5 10,0 12,5 15,0
Axial position, X 110

b 0,03 100 n=1

Average Nusselt number, Num

Present work
Re=100 n=1.5
Re=250 90 n=0.5
To p wa ll temper atu re, θ p

Re=500
0,02 80
Sung et al, [4] Re=100)
Re=250
Re=500 70
0,01 60

50
0,00
40

30
-0,01 -6 -5 -4 -3 -2 -1 0
0,0 2,5 5,0 7,5 10,0 12,5 15,0 10 10 10 10 10 10 10
Axial position, X Darcy number, Da

Fig. 4. Comparison of the local temperatures, (a) Bottom wall, (b) Top wall, with that Fig. 6. Effects of the index n on the variations of the average Nusselt number as
obtained by Sung et al. [4] for Da ¼ 105, S ¼ 0.5 and Rk ¼ 1. a function of the Darcy number for Re ¼ 100 and Rk ¼ 1.
1990 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995

1,0
n=0.5 n=0.5
n=0.5
n=1 n=1
0,8 n=1
n=1.5 n=1.5
n=1.5 -3
-5 Da=10
-4
Da=10
0,6 Da=10
x=5.75 x=5.75
Y x=5,75
0,4

0,2

a c e
0,0
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
U U U

1,0

0,8

0,6

Y n=0.5 n=0,5
n=0.5
0,4 n=1 n=1
n=1
n=1.5 n=1,5
n=1.5 -4 -3
-5 Da=10 Da=10
0,2 Da=10
x=6,75 b x=6.75 d x=6.75 f
0,0
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4

U U U

Fig. 7. Effects of Da and n on the velocity profiles at the midpoint of the porous blocks for Re ¼ 100 and type (B).

both hydrodynamic and thermal characteristics of the flow. The
fixed input parameters for all cases in the simulations are 3 ¼ 0.97,
CF ¼ 0.15 and Pr ¼ 10. The height and the width of the porous blocks
are both equal to 1 for type (A), and 0.5 for type (B).
4.1. Single porous block channel (type A)
The immediate effect of inserting a porous block which occupies
the height of the channel is the generation of an additional pressure
drop which is portrayed in terms of the normalized friction factor
fRe, in Fig. 5. As the permeability increases, fRe decreases rapidly.
This behavior is due to the diminution of the resistance to the flow
caused by the porous barrier. The additional pressure loss increases
with the power-law index and especially at low permeability which
corresponds to the Darcy regime. At the same shear rate, pseudo-
plastic fluids exhibit the lowest apparent viscosity. The gap exist-
ing between the fluids is augmented by the presence of the porous
substrate that creates an additional resistance to the flow which
increases when the permeability diminishes. Indeed, for Da ¼ 106,
the friction factor exhibited by dilatant fluids (n > 1) is approxi-
mately 1000 times the value obtained for pseudo plastic fluids
(n < 1). As Da increases, the difference in the pressure loss
exhibited by the fluids considered diminishes until it reaches an
asymptotic value corresponding to the non porous channel. This
reveals why using pseudo plastic fluids with porous media is
attractive.

a 1
0.85
0.65
0.5 0.45
0.25
Ψ = -0.1068 to 1 - 0.1 0.05
0
6 8 10 12 14 16 18

b 1
0.9 Ψ = -0.2536 to 1.048
0.7
0.5 0.5
- 0.9 0.3
0.1
0
6 8 10 12 14 16 18

c 1
0.9
0.7
Ψ = -0.067 to 1.022
0.5 0.5
0.3
-0.06 0.1
0
6 8 10 12 14 16 18

d e f
- 0.05 - 0.05
- 0.01

0.05 0.05 0.05

Fig. 8. Effects of (a) n ¼ 0.5, (b) n ¼ 1 and (c) n ¼ 1.5 on streamlines contours for Re ¼ 100 and Da ¼ 105. Enlarged views of the streamlines around the bottom wall porous block (d)
n ¼ 0.5, (e) n ¼ 1 and (f) n ¼ 1.5.
 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995 1991

a 1
0.7
0.9
0.5
0.5 0.3
0.1

0
-0.001 Ψ = -0.0018 to 1
6 8 10 12 14 16 18

1
b 0.7
0.9
0.5 0.5

- 0.15
0.3
0.1
Ψ = -0.189 to 1.017
0
6 8 10 12 14 16 18

1
c Ψ = -0.061 to 1.014 0.7
0.9
0.5 0.5
0.3
- 0.02 0.1
0
6 8 10 12 14 16 18

d e f
0.06 0.06 0.06

Fig. 9. Effects of (a) n ¼ 0.5, (b) n ¼ 1 and (c) n ¼ 1.5 on streamlines contours for Re ¼ 100 and Da ¼ 104. Enlarged views of the streamlines around the bottom wall porous block (d)
n ¼ 0.5, (e) n ¼ 1 and (f) n ¼ 1.5.

Concerning the thermal aspect, the variations of the average
Nusselt number, Num, versus the permeability are illustrated in
Fig. 6. It appears that a heat transfer enhancement is obtained over
the considered range of permeability. This is due to the fact that the
presence of the porous substrate makes both dynamic and thermal
boundary layers thinner. Moreover, as the power-law index
increases Num, decreases. This indicates that the combined use of
a porous matrix with a pseudo plastic fluid may lead to interesting
solutions for an effective heat transfer enhancement.
4.2. Two block porous channel (type B)
The geometrical layout considered makes the path of the fluid
through the channel sinuous and alters the appearance of the
velocity field. This is well illustrated in Fig. 7, where the combined
effects of the Darcy number and the index n on the stream-wise
velocity profiles at the middle of each of the porous blocks are
shown. For Da ¼ 105, Fig. 7a and b, the velocity practically
vanishes in the porous substrate, which exhibits a solid behavior.
This makes the flow concentrated in the unobstructed opening.
As result, a strong velocity gradient is created in the vicinity of
the bottom wall. As the permeability increases, the fluid pene-
trates better the porous blocks, leading to a decrease in the
maximum velocity (Fig. 7cef). Also notice that the pseudo plastic
fluid exhibit higher velocities near the wall. This points to
a vigorous convective activity near the heat sources. Such
behavior will have a significant impact on heat transfer as will be
shown later.

1
a
0.5

Ψ = 0 to 1
0
5 6 7 8 9 10 11 12

1
b Ψ = -0.011 to 1

0.5

0
5 6 7 8 9 10 11 12

1
c Ψ = -0.021 to 1

0.5

0
5 6 7 8 9 10 11 12

d e f

Fig. 10. Effects of (a) n ¼ 0.5, (b) n ¼ 1 and (c) n ¼ 1.5 on streamlines contours for Re ¼ 100 and Da ¼ 103. Enlarged views of the streamlines around the bottom wall porous block
(d) n ¼ 0.5, (e) n ¼ 1 and (f) n ¼ 1.5.
1992 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995

The stream function seems to be the most appropriate variable 250

to highlight how the flow field is altered by the presence of the
a
porous substrates. Fig. 8 is a typical representation of the mapping
200
flow obtained for Da ¼ 105, Re ¼ 100 and n ¼ 0.5, 1 and 1.5. A close
examination of this figure reveals certain distinct features and
Re=300
trends. First, due to the low permeability, the main flow is 150 Da=10
-3

concentrated in the free passage between the channel walls and the -4
Da=10
porous blocks, so that there is virtually no penetration of fluid in Nux 100 n=1.5
Re=100
these blocks which may then be regarded as solid blocks (as shown
in Fig. 7def). Thus, the flow area is reduced, leading to a narrowing
of the streamlines near the channel walls, and, thereby intensifying 50
the convective effects. Second, due to the interaction of the flow
and the porous blocks, five vortex effects are generated: a weak 0
recirculation above each porous block, a vortex downstream each
block and an eddy zone attached to the upper block. This complex
-50
flow field is due to the sinuous flow path, the suction effect caused 5,0 5,5 6,0 6,5 7,0 7,5
by the adverse pressure below each block and the separation and x
reattachment of the boundary layer. The extent and intensity of the
recirculation between the blocks is slightly affected by the index n. 500
However, the recirculation behind the second block is very sensi-
b
tive to the variation of n, and particularly as regards its size which
400
increases dramatically with the decrease of n. As Da increases, the
resistance to the flow due to the porous medium diminishes and Re=300
-3
the flow rate that crosses the porous barriers augments. This leads 300 Da=10
-4
to the decrease in size and intensity of the recirculation zones as Da=10
shown in Figs. 9 and 10. Nux 200 n=0.5
Re=100
The normalized pressure drop for type (A) and type (B) is
sketched in Fig. 11 against the Darcy number for different values of
n. For the type (B), the Newtonian and dilatant fluids show almost 100
the same friction loss for Da below approximately 103. In this
second configuration, the pseudo plastic fluids (n < 1), also, exhibit 0
the lowest resistance to the flow. It appears, clearly, that the type
(B) reduces substantially the pressure drop. Thus, for example, for
Da ¼ 106, the dimensionless friction coefficient is divided -100
5,0 5,5 6,0 6,5 7,0 7,5
approximately by 23,000 for the dilatant fluid (n > 1), by 660 for
the Newtonian fluids and by 20 for the pseudo-plastic fluids (n < 1). x
It is reasonable to expect that the change in the nature of the
Fig. 12. Effects of Re and Da on the local Nusselt number evolution for Pr ¼ 10 and
flow field due to the porous inserts will strongly affect the thermal Rk ¼ 1.
behavior of the channel. Fig. 12 shows the distribution of the local
Nusselt number and highlights the combined effects of the Rey-
nolds number, the Darcy number and the index n. From fluid (n ¼ 0.5). This makes this kind of fluids (n < 1) thermally
a phenomenological point of view, Nux displays variations which more interesting for type (B) as well. The sudden increase of Nux to
are not affected by the fluid rheology. However, the exhibited its maximum value at the entrance of the first block is due mainly to
values are significantly more important with the pseudo-plastic two phenomena. First, this position coincides with the start of
heating which gives rise to a thermal boundary layer characterized
by large parietal temperature gradients. Second, the presence of the
porous substrate lead to the narrowing of the flow passage section,
10
6
causing an acceleration of the fluid near the lower wall as was
Average friction coeficient, fRem

n=0,5
previously shown in Figs. 7 and 8. From this position, Nux decreases
n=1
10
5 as the fluid crosses the channel. This happens because the heat
n=1,5
transfer rate is slowed down with the development of the hydro-
10
4 dynamic and thermal boundary layers. For the second porous block,
the same thermal behavior is observed with lower intensity,
10
3 because the fluid has accumulated a certain amount of heat during
its passage through the first block, thus increasing its temperature.
2 In addition, the heat transfer is enhanced by increasing the Rey-
10
nolds number. This is due to a higher velocity near the channel wall
1 for larger Reynolds number values, which enhances the thermal
10
energy transport capacity of the fluid. While increasing the
0
permeability decreases Nux in the first block, the opposite effect is
10
-6 -5 -4 -3 -2 -1 0 observed in the second substrate. This can be explained by the fact
10 10 10 10 10 10 10
that when Da increases, the fluid penetrates better in the porous
Darcy number, Da
substrates as shown in Figs. 8e10.
Fig. 11. Comparison of the averages friction coefficients with the Darcy number for Fig. 13 illustrates the variation of the average Nusselt number
different values of n of the two configurations and Re ¼ 100. against the Darcy number for different values of the Reynolds
 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995 1993

160 observe that as n increases, the average Nusselt number decreases.

This appears to be very interesting according to the effect of the
n=0.5
power-law index on the pressure drop (Fig. 11). As expected,
n=1.5
increasing Re significantly enhances the heat transfer. However, the
n=1
120 pressure drop is also increased. At last, one can note that the pseudo
plastic fluids (n < 1) allow to obtain almost the same heat transfer
Re=300 for Re ¼ 100 than that obtained with dilatant fluids with Re ¼ 300.
Num Fig. 14 displays the impact of the porous media and the non-
Newtonian fluid on the local bottom wall temperature. For
80
Da ¼ 106, and in the second block region, the wall temperature
exhibits higher important values comparatively to those observed
in the first block region. Indeed, for this permeability value the
highest value of qw in the second block is approximately 5.55 times
40 Re=100 more important than the value obtained in the first bloc for n ¼ 0.5.
-6 -5 -4 -3 -2 -1 0
qw is 5.42 times larger for n ¼ 1 and 5.25 times larger for n ¼ 1.5. The
10 10 10 10 10 10 10 increase of the permeability leads to an important reduction in qw
Da in the second bloc while on the first bloc, qw augments. This is due
to the resistance opposed to the flow by the porous matrices which
Fig. 13. Effects of the Reynolds number and the index n on the variations of the decreases as Da increases. This makes more fluid penetrates in the
average Nusselt number for Rk ¼ 1.
porous blocks, thus enhancing the convective effects and reducing
them on the second bloc. Moreover, the lowest temperatures are
number and the index n. The thermal behavior shown on these obtained for intermediate permeability values (Da ¼ 102 and 104)
curves is identical for the three considered fluids. Unlike for type for all values of the index n, while as n decreases the wall
(A), Num is not a monotonous function of Da anymore. Thus, the temperature decreases.
Num increases when the porous layer permeability increases up to The comparison of the thermal performances of type (A) and
a critical value beyond which, Num decreases until it reaches an type (B) is shown in Fig. 15. Overall, the first configuration provides
asymptotic value, which corresponds to the non-porous case. This the best heat transfer due to a larger quantity of porous material
result is obtained for an intermediate permeability for which the used. However, the difference in Num, diminishes with increasing
pressure loss is lower. Concerning the rheological aspect, one can the porous blocks permeability. Notice that there is a particular

a b 0,28
0,25 Da=1 Da=1
-2
Da=10 0,24 Da=10
-2
-4
0,20 Da=10 Da=10
-4
-6
Da=10 0,20 -6
Da=10
n=0,5 n=1
0,15 0,16
θp θp
0,12
0,10
0,08
0,05
0,04

0,00 0,00

5,0 5,5 6,0 6,5 7,0 7,5 8,0 8,5 5,0 5,5 6,0 6,5 7,0 7,5 8,0 8,5

x x
c d
0,24 0,14
Da=1 n=1
Da=10
-2
0,12 n=1,5
0,20 -4 n=0,5
Da=10 -4
-6 0,10 Da=10
Da=10
0,16
n=1,5 0,08

θ p 0,12 θp 0,06

0,08 0,04

0,02
0,04
0,00
0,00
-0,02
5,0 5,5 6,0 6,5 7,0 7,5 8,0 8,5 5,0 5,5 6,0 6,5 7,0 7,5 8,0 8,5

x x
Fig. 14. Effects of Da and n on the variations of the bottom wall local temperature for Re ¼ 100 and Rk ¼ 1.
1994 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995

with a single porous block and one with two porous inserts
100
n=0,5 mounted alternatively on the upper and lower walls of the channel.
90 n=1 The effect on the flow pattern and the heat transfer of several
n=1,5 parameters such as the Darcy number, the Reynolds number, the
80 power-law index and the thermal conductivity ratio are analyzed. It
is found that pseudo plastic fluids generate the highest heat
70 transfer with the lowest pressure drop in both configurations. The
Num results show that, from a thermal point of view, type (A) is more
60
advantageous, while from a dynamic point of view, type (B) is
50 better. Throughout this investigation, some interesting results are
revealed, showing that the thermal efficiency of type (B) can be
40 improved in several ways. This is true even if a less porous matter is
used. First, type (B) can be used with a pseudo plastic fluid instead
30
of the Newtonian fluid. Second, as the pressure loss is lower for this
-6 -5 -4 -3 -2 -1 0
10 10 10 10 10 10 10 configuration, the Reynolds number can be increased to improve
heat transfer while maintaining the pumping power at reasonable
Da levels. Third, using porous inserts with higher thermal conduc-
tivity, the heat transfer can be improved without increasing the
Fig. 15. Comparison of the variations of the average Nusselt numbers of the two
configurations with the Darcy number for different n, Re ¼ 100 and Rk ¼ 1.
pressure drop.

Rk=1 References
800
Rk=5
[1] A. Hadim, Forced convection in porous channel with localized heat sources,
Rk=10
J. Heat Transfer 116 (1994) 465e471.
600 Rk=15 [2] M. Hamdan, M.A. Al-Nimr, The use of porous fins for heat transfer augmen-
tation in parallel-plate channels, Transp. Porous Med. (2009). doi:10.1007/
Rk=20 s11242-009-9510-2.
Rk=50 [3] P.C. Huang, K. Vafai, Analysis of forced convection enhancement in a channel
400 using porous blocks, J. Thermophys. Heat Transfer 8 (1994) 563e573.
Nu x Rk=75 [4] H.J. Sung, S.Y. Kim, M.J. Hyun, Forced convection from an isolated heat source
Rk=100 in a channel with porous medium, Int. J. Heat Fluid Flow 16 (1995) 527e535.
[5] T.W. Tong, M.C. Sharatchandra, Z. Gdoura, Using porous inserts to enhance
200 n=0.5 heat transfer in laminar fully-developed flows, Int. Commun. Heat Mass
Transfer 20 (1993) 761e770.
[6] S. Chikh, A. Boumedien, K. Bouhadef, G. Lauriat, Analytical solution of non-
Darcian forced convection in an annular duct partially filled with a porous
0
medium, Int. J. Heat Mass Transfer 38 (1995) 1543e1551.
[7] D.A. Nield, A.V. Kuznetsov, The effect of a transition layer between a fluid and
5,0 5,5 6,0 6,5 7,0 7,5 a porous medium: shear flow in a channel, Transp. Porous Med. 78 (2009)
x 477e487.
[8] Y. Ould-Amer, S. Chikh, K. Bouhadef, G. Lauriat, Forced convection cooling
enhancement by use of porous materials, Int. J. Heat Fluid Flow 19 (1998)
Fig. 16. Effects of the thermal conductivities ratio, Rk, on the local Nusselt number for 251e258.
n ¼ 0.5, Re ¼ 100 and Da ¼ 105. [9] S. Chikh, A. Boumedien, K. Bouhadef, G. Lauriat, Analysis of fluid flow and heat
transfer in a channel with intermittent heated porous blocks, Heat Mass
Transfer 33 (1998) 405e413.
[10] M. Alkam, M.A. Al-Nimr, M. Hamdan, Enhancing heat transfer in parallel-plate
value of Da, depending on n, beyond which the type (B) dissipates
channels by using porous inserts, Int. J. Heat Mass Transfer 44 (2001)
heat better. In addition, for Day2:104, the pseudo-plastic fluid in 931e938.
type (B), provides a heat transfer similar to that obtained with [11] B.M. Miranda, N.K. Anand, Convective heat transfer in a channel with porous
Newtonian and dilatant fluids (n > 1) with type (A), while the baffles, Numer. Heat Transfer Part A 46 (2004) 425e452.
[12] P.C. Huang, C.F. Yang, J.J. Hwang, M.T. Chiu, Enhancement of forced-
exhibited pressure drop is reduced as it was already shown in Fig. 5. convection cooling of multiple heated blocks in a channel using porous
The thermal conductivity of a porous medium is an important covers, Int. J. Heat Mass Transfer 48 (2005) 647e664.
parameter that significantly influences the heat transfer. Then, to [13] H.Y. Li, K.C. Leong, L.W. Jin, J.C. Chai, Analysis of fluid flow and heat transfer in
a channel with staggered porous blocks, Int. J. Therm. Sci. 49 (2010) 950e962.
improve the thermal efficiency of type (B) the thermal conductivity [14] Z.F. Huang, A. Nakayama, K. Yang, C. Yang, W. Liu, Enhancing heat transfer in
ratio Rk is varied from 1 to 100. A typical result is illustrated in the core flow by using porous medium insert in a tube, Int. J. Heat Mass
Fig. 16 for the pseudo plastic fluid. One can note the significant Transfer 53 (2010) 1164e1174.
[15] P.C. Huang, C.F. Yang, Analysis of pulsating convection from two heat sources
effect of Rk on Nux in the second block while in the first block the mounted with porous blocks, Int. J. Heat Mass Transfer 51 (2008) 6294e6311.
effect vanishes. This result is due to the particular arrangement of [16] V.V. Satyamurty, D. Bhargavi, Forced convection in thermally developing
the blocks. Thus, it is sufficient to use high conductivity porous region of a channel partially filled with a porous material and optimal porous
fraction, Int. J. Therm. Sci. 49 (2010) 319e332.
matter only in the second block to enhance heat transfer. Moreover,
[17] G. Chen, H.A. Hadim, Numerical study of non-Darcy forced convection in
this enhancement is obtained without any augmentation of the a packed bed saturated with a power-law fluid, J. Porous Med. 1 (1998)
pressure drop. 147e157.
[18] M.K. Alkam, M. Al-Nimr, Z. Moussa, Forced convection of non-Newtonian
fluids in porous concentric annuli, Int. J. Numer. Meth. Heat Fluid Flow 8
(1998) 703e716.
5. Conclusion [19] M. Al-Nimr, T.K. Aldoss, The effect of the macroscopic local inertial term on
the non-Newtonian fluid flow in channels filled with porous media, Int. J. Heat
A detailed numerical analysis is presented for heat transfer Mass Transfer 47 (2004) 125e133.
[20] R. Nebbali, K. Bouhadef, Numerical study of forced convection in a 3D flow of
enhancement in a channel by using porous blocks and non- a non-Newtonian fluid through a porous duct, Int. J. Numer. Meth. Heat Fluid
Newtonian fluids. Two channel configurations are considered, one Flow 16 (2006) 870e889.
 R. Nebbali, K. Bouhadef / International Journal of Thermal Sciences 50 (2011) 1984e1995 1995

[21] Y. Aksoy, M. Pakdemirli, Approximate analytical solutions for flow of a third-
grade fluid through a parallel-plate channel filled with a porous medium,
Transp. Porous Med. 83 (2010) 375e395.
[22] Y.L. Chen, X.D. Cao, K.Q. Zhu, A gray lattice Boltzmann model for power-law
fluid and its application in the study of slip velocity at porous interface,
J. Non-Newtonian Fluid Mech. 159 (2009) 130e136.
[23] A.V. Shenoy, Non-Newtonian fluid heat transfer in porous media, Adv. Heat
Transfer 24 (1994) 101e190.
[24] R.H. Christopher, S. Middleman, Power-law flow through a packed tube, Ind.
Eng. Chem. Fund. 4 (1965) 422e426.
[25] R.V. Dharmadhikari, D. D Kale, Flow of non-Newtonian fluids through porous
media, Chem. Eng. Sci. 40 (1985) 527e529.
[26] S.V. Patankar, Numerical Heat Transfer and Fluid Flow. Hemisphere, New
York, 1980.
[27] B. Zhang, Y. Zhao, A numerical method for simulation of forced convection in
a composite porous fluid system, Int. J. Heat Fluid Flow 21 (2000) 432e441.