International Journal of Heat and Mass Transfer 123 (2018) 600–615

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Buoyancy driven double diffusive moisture convection inside the

fluid-porous-solid sandwiched building enclosure containing internal
heating sources
Jiang-Tao Hu a,b,c, Shuo-Jun Mei a,b,c, Di Liu d, Fu-Yun Zhao a,b,c,⇑, Han-Qing Wang e
a
Key Laboratory of Hydraulic Machinery Transients (Wuhan University), Ministry of Education, Wuhan, Hubei Province, China
b
Hubei Key Laboratory of Waterjet Theory and New Technology (Wuhan University), Wuhan, Hubei Province, China
c
School of Power and Mechanical Engineering, Wuhan University, Wuhan, Hubei Province, China
d
College of Pipeline and Civil Engineering, China University of Petroleum, Qingdao, Shandong Province, China
e
School of Civil Engineering, University of South China, Hengyang, Hunan Province, China

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

Article history: Double diffusive convection in an enclosure partially inserted with heat-generated porous layer and solid
Received 3 October 2017 layer is investigated numerically. Moist air motions are simultaneously driven by the internal heat gen-
Received in revised form 30 December 2017 eration and external concentration difference imposed across the enclosure. Identical porous and solid
Accepted 18 February 2018
layers are attached to the vertical walls, and the walls are held at constant temperatures and concentra-
Available online 20 March 2018
tions. The horizontal walls of the enclosure are assumed to be adiabatic and impermeable. The aspect
ratio of the cavity is equal to 1.0 and the saturating fluid is moist air (Pr = 0.7 and Le = 0.8). Present
Keywords:
research covers the range of Rat from 103 to 107, N from
10.0 to 10.0, Da from 10
9 to 10
1, U from
Double diffusive convection
Thermal storage composite wall

90° to 90°, and Krf and Krs from 0.1 to 10.0. The results show that the thermal and solutal buoyancy
Thermal conductivity ratio forces are always aiding with each other and the minus sign of N only has influence on the fluid orien-
Internal heat generation tations. For the vertical case, heat and moisture transfer of fluid/porous interface is analyzed as a function
of the buoyancy ratio. Average Nusselt and Sherwood numbers are increasing functions of Rat, N, Krp and
Krs. Furthermore, volume averaged heat transfer rate as functions of inclination angle, for N > 0.0, is pre-
sented like a sinusoid, where the maximum is near at U = 45° and 135°, whereas the minimum is located
at U = 90°. Present research could benefit future development of sustainable building energy storage.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction composite-wall, which consists of porous and solid layers, may

With the rapid depletion of the fossil fuels and the increase of
global energy demand, solar energy has been considered as an
effective alternative by many researchers as the earth surface
receives abundant solar energy throughout the whole calendar
year. The thermal storage wall has been used more extensively in
buildings to store the solar energy [1]. In general, thermal energy
storage system consists of porous materials, such as alumina and
zeolite that are used for dehumidification and removal of pollutant
gases from air [2,3]. However, thermal storage wall has the draw-
back of low thermal resistance, which leads to heat loss at night.
Furthermore, as the energy storage and supply cannot be con-
trolled, there is a possibility of overheating in the heated room. A
⇑ Corresponding author at: School of Power and Mechanical Engineering, Wuhan
University, Dong-Hu Southern Road, 430072 Wuhan, Hubei Province, China.
E-mail addresses: liudi@upc.edu.cn (D. Liu), fyzhao@whu.edu.cn (F.-Y. Zhao).
overcome this drawback.
Many works in this topic concern thermal natural convection in
enclosures containing simultaneously a fluid reservoir and a por-
ous layer saturated with the same fluid. Singh et al. [4] investigated
a composite cavity containing a fluid layer overlying a porous layer
saturated with the same fluid. The amount of fluid penetration into
the porous layer depends strongly upon the Darcy, thermal and
solutal Rayleigh numbers and heat and mass transfer on the heated
wall neat the interface depends strongly on the Darcy number.
Mharzi et al. [5] made a numerical study of natural thermosolutal
convection in an elongated enclosure, partitioned by a vertical por-
ous layer, which observed that the heat and mass transfer were
sensitive to the thermal conductivity ratio and solutal diffusivity
ratio. Zhao et al. [6] carried out linear stability analysis to predict
the onset of double-diffusive convection in superposed fluid and
porous layer using a one-equation model. Results illuminated dif-
ferences in the critical conditions and the flow streamlines at onset
compared with the same problems by Chen et al. [7] using a

https://doi.org/10.1016/j.ijheatmasstransfer.2018.02.071
0017-9310/Ó 2018 Elsevier Ltd. All rights reserved.
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
Nomenclature
AR enclosure aspect ratio
d thickness of porous and solid layers
D mass diffusivity
Da Darcy number
g gravitational acceleration
H width of the enclosure
k thermal conductivity
K permeability of porous medium
Kr thermal conductivity ratio
Kc mass diffusion coefficient ratio
L length of the enclosure
Le Lewis number
N buoyancy ratio
Nu average Nusselt number
P dimensionless pressure
Pr Prandtl number
q00 strength of internal heat generation
RaT thermal Rayleigh number
Sh average Sherwood number
s concentration of species
S dimensionless concentration
DS dimensionless concentration difference
t temperature
T dimensionless temperature
two-equation model. The influence of isotropic or anisotropic por-
ous layers for natural convection driven by thermal and solutal
buoyancy forces in an enclosure has numerically analyzed by Ben-
nacer et al. [8]. They obtained an optimum value of Nusselt num-
ber, which was a function of the anisotropic parameter. Gobin
et al. [9,10] dealt with the similar problem, where the coupling
of the flow penetration in the porous layer with combined buoy-
ancy forces induces a specific behavior of the flow structure and
average heat transfer in the enclosure. Baytas et al. [11] investi-
gated the double diffusive natural convection between a saturated
porous layer and an overlying fluid layer. The results show that the
height of the step at the interface between fluid and porous layer
has a significant effect on the flow field and heat and mass transfer
from the left-hand to the right-hand walls. The conjugate double-
diffusive natural convective flows in a variable porosity layer sand-
wiched between two walls has been investigated by Al-Farhany
et al. [12], which presented results for a variable porosity medium
with different values of the non-dimensional governing parame-
ters. Hadidi et al. [13–15] extended previously published numeri-
cal studies on double diffusive convection in partially porous 2D
enclosures to three dimensions in cubic enclosures, which dis-
cussed the effect of the third direction on the flow, heat and mass
transfer in comparison with the simplified (2D) approach. A cavity
that has a fluid and superposed porous medium with an adiabatic
rotating cylinder was studied by Ismael et al. [16], which showed
that Rayleigh number, Darcy number and cylinder size could have
profound effects on the local and averaged heat transfer.
In addition, solid partition is also considered by many
researchers due to inhibition effect in heat and mass transfer.
The study of natural convection with radiation in a rectangular
enclosure divided by two partitions was numerically examined
by Han et al. [17], whom analyzed the influence of the partition
configuration in radiation-affected steady-laminar natural con-
vection. Zhao et al. [18] investigated the comparison of vertical
enclosures surrounded by thermal diffusive walls and ones
inserted with a solid heat conducting partition. A relative low
thermal conductivity partition could enhance the heat transfer
rate across the enclosure of moderate volume ratio scale. Sankha-
601
DT dimensionless temperature difference
u, v velocity components in x, y direction
U, V dimensionless velocity components
x, y Cartesian coordinates
X, Y dimensionless Cartesian coordinates
Greek symbols
a thermal diffusivity
bt coefficient of thermal expansion
bc coefficient of solutal expansion
e porosity
l dynamic viscosity
m kinematic viscosity
q density
U angle of the inclination
Subscripts
0 reference state
c low level
f fluid
h higher level
p porous
s solid
vara et al. [19] considered the influence of partition length on
natural convection in a partially divided square enclosure, which
observed that the Nusselt number could decrease with increasing
partition length from 0 to 0.5. The research of heat and moisture
transfer by conjugate natural convection in a partial enclosure
with a solid wall was considered by Liu et al. [20], their results
showed that heat transfer potential, mass transfer potential,
and volume flow rate can be promoted or inhibited, depending
strongly on the wall materials and size, thermal and moisture
Rayleigh numbers. Costa [21] has investigated the effects of ther-
mal boundary conditions and thermal conductivity of the parti-
tions on the conjugate natural convection. His research has
demonstrated that some different combinations of the placement
and length of the partitions could lead to the same thermal per-
formance of the enclosure, the best option being that correspond-
ing to the shorter partitions, and thus to the lower cost of the
material used in the partitions. Harzallah et al. [22] performed
a vertical enclosure bounded by finite thickness walls with
opposing temperature, concentration gradients on vertical walls,
which showed that the fluid conduction dominates for higher
fluid-to-solid thermal conductivity ratio. Chamkha et al. [23–26]
dealt with the problem of natural convection in porous enclo-
sures bounded with a solid wall, which observed an uncommon
behavior of the heat transfer with the wall thickness. Hu et al.
[27,28] analytically and numerically studied enclosures contain-
ing many solid obstacles. It was founded that the presence of
solid obstacles suppresses the intensity of convection.
The literature reviews indicate that most of former researches
devoted to single or multi-component fluid, such as single compo-
nent conjugate heat transfer [29,30], internal heat generation fluid
[31–34], power-law fluids [35,36], nanofluids [27,37,38], micropo-
lar fluid [39] or porous media [40–46]. However, it can be observed
that little information has been obtained for double diffusive con-
vection inside an enclosure with the thermal storage composite
wall. The present work is motivated by a need to understanding
the heat and moisture transfer in building structures with the ther-
mal storage composite wall, to provide heating engineers with an
effective design tool against heat loss.
602 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615

Table 1
Average Nusselt and Sherwood numbers compared with those of Singh et al. [4],
Baytas et al. [11] and Hadidi et al. [15] (Darcy model: N = 0, AR = 1, Ra = 105).

0.84 1.05 Le
10
6 10
5 10
6 Da
Singh et al. [4] 4.096 4.113 3.997 Nu
3.710 3.725 4.126 Sh
Baytas et al. [11] 4.42 4.50 4.39 Nu
4.0 4.01 4.44 Sh
Hadidi et al. [15] 4.41 4.42 4.29 Nu
4.02 4.03 4.39 Sh
Present work 4.283 4.385 4.356 Nu
4.061 4.072 4.395 Sh

In this paper, the major features of steady-state double diffusive

Fig. 1. Schematic of the enclosure with one (a) and two (b) porous layer.
convection in a square enclosure, partially filled with a porous
layer with spatially uniform internal heat generation and a solid
layer, are to be explored. Following that, detailed parametric inves-
tigations will be conducted, regarding of different governing
parameters. For some relevant situations, the flow, temperature
and concentration fields are also presented and analyzed by the
streamlines, isotherms and iso-concentrations. Finally, conclusions
will be drawn for benefiting future thermal storage walls design
and relevant fields.

Isotherms

Streamlines

Iso-concentrations

Fig. 2. Comparison between present study and Singh et al. [4] for RaT = 105, N = 10, Da = 10
3 and Le = 0.84.
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615 603

2. Problem statements and mathematical formulations layers. It is assumed that the dimension in the z direction is large
enough and the end effects on the flow are negligible, i.e., fluid flow
The physical domain under investigation is, as illustrated in and heat transfer is two-dimensional. Horizontal walls of enclo-
Fig. 1, an inclined square enclosure bounded with porous and solid sures are assumed to be impermeable and perfect thermal insula-

Table 2
Validation of the numerical code in the case of double diffusive convection for Le = 0.8, N = 1, Ra = 105 and various the thickness of solid partition d, comparing with those of Jiang-
tao Hu [27].

d Nu Sh
Jiang-tao Hu [27] Ours Jiang-tao Hu [27] Ours
0 5.2456 5.2461 4.6785 4.6791
0.05 4.0479 4.0481 3.7628 3.7684
0.1 3.2702 3.2751 3.1169 3.1201

Table 3
The grid independency test for Nusselt number (Pr = 0.71, Le = 0.8, Ra = 106, Da = 10
3, Krp = Krs = 1.0 and U = 0°).

N 80  80 100  100 120  120 150  150 200  200

0 0.3602 0.3675 0.3776 0.3751 0.3695
10 0.8862 0.8750 0.8568 0.8494 0.8501
(a) Rat = 103
(b) Rat = 105
(c) Rat = 106
(d) Rat = 107

Streamlines Isotherms Iso-concentrations

Fig. 3. Contour lines of stream function, temperature, and concentration for Da = 10
3, N = 1.0, Krp = Krs = 1.0, and U = 0°, (a) hmax = 0.0522; (b) hmax = 0.0425; (c) hmax =
0.0249; (d) hmax = 0. 0188.
604 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615

tions, while constant temperature and different uniform concen- porous regions, with proper interpretations of the physical
trations are maintained at the vertical walls of the enclosure. The variable.
enclosure is of height H and width L and is titled at an angle with Two dimensional conservation transport equations of continu-
respect to the horizontal ground. The porous and solid layers are of ity, momentum, energy and mass are described respectively as
equal width d and placed in the right vertical wall. A uniformly dis- follows,
tributed internal heat generation of constant strength q00 exists in
the porous medium. The thermal conductivities of the porous @u @ v
þ ¼0 ð1Þ
and solid layers are assumed kp and ks, respectively. @x @y

Fluid layer,
2.1. Governing equations !
@u @u 1 @p @2u @2u
u þv ¼
þv þ þ gbt ðt
tc Þ sin U
Cavities in enclosure are filled with the Newton-Fourier fluid of @x @y q0 @x @x2 @y2
thermal conductivity kf and mass diffusivity Df, which is incom-
þ gbc ðs
sc Þ sin U ð2Þ
pressible and follows the Boussinesq approximation. Both the vis-
cous dissipation and Soret and Dufour effects are neglected. No !
chemical reaction is assumed exit and the fluid is assumed to be @v @v 1 @p @2v @2v
u þv ¼
þv þ þ gbt ðt
tc Þ cos U
radiatively nonparticipating so that the thermal radiation transfer @x @y q0 @y @x2 @y2
between the interior faces of both enclosures is negligible. It is
assumed that the flow field within enclosure occurs only in the þ gbc ðs
sc Þ cos U ð3Þ
laminar regime. For the porous region, the Brinkman-extended !
Darcy model is adopted and the porous medium is assumed to @t @t @2t @2t
u þv ¼ af þ ð4Þ
be homogeneous, isotropic and saturated with moist air. In the @x @y @x2 @y2
present approach, a single set equation is used for both fluid and
(a) N = 0.0
(b) N = -1.0
(c) N = -10.0
(d) N = 10.0

Streamlines Isotherms Iso-concentrations

Fig. 4. Contour lines of stream function, temperature, and concentration for Rat = 106, Da = 10
3, Krp = Krs = 1.0 and U = 0°, (a) hmax = 0.046; (b) hmax = 0.0244; (c) hmax =
0.0183; (d) hmax = 0.0184.
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615 605

! !
@s @s @2s @2s @T @T Kr p
1 @ 2 T @ 2 T Kr p
u þv ¼ Df þ ð5Þ U þV ¼ pffiffiffiffiffiffiffiffiffiffi þ þ pffiffiffiffiffiffiffiffiffiffi ð20Þ
@x @y @x2 @y2 @X @Y RaPr @X 2 @Y 2 RaPr

Porous layer,
!
  !
@S @S Kcp @2S @2S
1 @u @u 1 @p @2u @2u v U þV ¼ pffiffiffiffiffiffiffiffiffiffi þ ð21Þ
u þv ¼
þv þ
v @X @Y Le RaPr @X 2 @Y 2
e2 @x @y q0 @x @x2 @y2 K
þ gbt ðt
t c Þ sin U þ gbc ðs
sc Þ sin U ð6Þ Solid layer,

  !     
1 @v @v 1 @p @2v @2v v @ @T @ @T
u þ v ¼
þ v þ
v þ ¼0 ð22Þ
e2 @x @y q0 @y @x2 @y2 K @X @X @Y @Y

þ gbt ðt
t c Þ cos U þ gbc ðs
sc Þ cos U ð7Þ     

@ @S @ @S
! þ ¼0 ð23Þ
@X @X @Y @Y
@t @t @2t @2t q00
u þv ¼ ap þ þ ð8Þ
@x @y @x2 @y2 q0 C p
where Krp and Krs respectively represent the ratios of the thermal
! conductivities kp and ks to kf; Kcp and Kcs are the ratio of mass trans-
@s @s @2s @2s fer coefficients qpDp and qsDs to qoDf, respectively. The effective vis-
u þv ¼ Dp þ ð9Þ
@x @y @x2 @y2 cosity of porous and solid medium mp and ms are assumed to be equal
to the fluid one mf. The analysis is performed for high porosity mate-
Solid layer, rial and the porosity of the solid matrix appears only in the inertia
    
@ @t @ @t term. In this case, the inertia term is relatively negligible in porous
þ ¼0 ð10Þ media. This reason allows us to set the porosity e equal to one. All
@x @x @y @y
definitions are given in the nomenclature.
    
@ @s @ @s
þ ¼0 ð11Þ
@x @x @y @y
It is convenient to non-dimensionalize Eqs. (1)–(11) by using
the following relations,
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðX; YÞ ¼ ðx; yÞ=H; ðU; VÞ ¼ ðu; v Þ= gbDtH;
P ¼ p=q0 gbDtH; T ¼ ðt
t 0 Þ=Dt; Dt ¼ q00 H2 =kp ;
S ¼ ðs
s0 Þ=DS; DS ¼ sh
sc ð12Þ
Hence, the non-dimensional governing equations are,
@U @V
þ ¼0 ð13Þ
@X @Y
Fluid layer,
rffiffiffiffiffiffi 2 !
@U @U @P Pr @ U @ 2 U
U þV ¼
þ þ þ ðT þ NSÞ sin U ð14Þ
@X @Y @X Ra @X 2 @Y 2

rffiffiffiffiffiffi 2 !
(a)
@V @V @P Pr @ V @ 2 V
U þV ¼
þ þ þ ðT þ NSÞ cos U ð15Þ
@X @Y @Y Ra @X 2 @Y 2

!
@T @T 1 @2T @2T
U þV ¼ pffiffiffiffiffiffiffiffiffiffi þ ð16Þ
@X @Y RaPr @X 2 @Y 2
!
@S @S 1 @2S @2S
U þV ¼ pffiffiffiffiffiffiffiffiffiffi þ ð17Þ
@X @Y Le RaPr @X 2 @Y 2

Porous layer,
  rffiffiffiffiffiffi 2 ! rffiffiffiffiffiffi
1 @U @U @P Pr @ U @ 2 U Pr 1
U þ V ¼
þ þ
U
e2 @X @Y @X Ra @X 2 @Y 2 Ra Da
þ ðT þ NSÞ sin U ð18Þ

  rffiffiffiffiffiffi 2 ! rffiffiffiffiffiffi
1 @V @V @P Pr @ V @ 2 V Pr 1
U þ V ¼
þ þ
V (b)
e2 @X @Y @Y Ra @X 2 @Y 2 Ra Da
Fig. 5. Effect of thermal Rayleigh number and buoyancy ratio on heat and mass
þ ðT þ NSÞ cos U ð19Þ
transfer for Da = 10
3 and Krp = Krs = 1.0, (a) Nusselt number, (b) Sherwood number.
606 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615

Here, aspect ratio of the enclosure AR, Prandtl number Pr, Lewis X ¼ 1; T ¼ 0; S¼0 ð28Þ
number Le, Rayleigh number Ra, Darcy number Da and buoyancy
The problem under analysis is a conjugate heat and mass trans-
ratio N could be defined as follows,
fer problem where it is assumed that, at each interface fluid-porous
H m a gbDtH3 and porous-solid wall of the enclosure,
AR ¼ ; Pr ¼ ; Le ¼ ; Ra ¼ ;
L a D ma @T f @T p @Sf @Sp
K bs Ds X ¼ 1
2d; ¼ Kr p ; ¼ Kcp ð29Þ
Da ¼ ; N¼ ð24Þ @Y @Y @Y @Y
H 2 bt Dt
@T f @T s @Sf @Ss
X ¼ 1
d; ¼ Kr s ; ¼ Kcs ð30Þ
2.2. Boundary conditions and convective transport @Y @Y @Y @Y
The average heat and mass transfer rates of interface between
Non-slip boundary conditions and continuities of temperature fluid and porous regions are used to describe the heat and mass
and concentration are imposed over the porous and solid wall transfer characteristics and thus are of interest in engineering
and boundaries of the enclosure, applications, and they could be expressed in terms of Nusselt
W ¼ U ¼ V ¼ 0; Tf ¼ Tp ¼ TS ð25Þ and Sherwood numbers respectively. It should be noted that the
enclosure is inserted with one porous wall with spatially uniform
At the impermeable and perfect thermal insulated horizontal internal heat generation. Therefore, Nusselt and Sherwood num-
walls, bers are defined as [48],
@T @S Z Z
Y ¼ 0; AR; ¼ ¼0 ð26Þ AR
q00 AR
Dt Kr p
@X @X Nu ¼ dY ¼ Kr p dY ¼ R AR
2
Constant temperature and concentration at the vertical walls, 0 ktX¼1
2d =H 0 tX¼1
2d
0
T X¼1
2d dY

X ¼ 0; T ¼ 0; S¼1 ð27Þ ð31Þ

(a) Da = 10-7
(b) Da = 10-5
(c) Da = 10-4
(d) Da = 10-1

Streamlines Isotherms Iso-concentrations

Fig. 6. Contour lines of stream function, temperature, and concentration for Rat = 106, N = 1.0, Krp = Krs = 1.0, and U = 0°, (a) hmax = 0.0364; (b) hmax = 0.0366; (c) hmax = 0.0335;
(d) hmax = 0.0246.
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615 607

Z AR  pffiffiffiffiffiffiffiffiffiffi  
@S  First, the case of a vertical enclosure, i.e. U = 0, is presented in
Sh ¼ Le PrRaUS
Kcp  dY ð32Þ
0 @X X¼1
2d detail for the range of Ra from 103 to 107, N from
10 to 10, Da
from 10
1 to 10
9, Krp from 0.1 to 10 and Krs from 0.1 to 10. Then
where X1-2d is the interface of fluid/porous. the effect of the angle of inclination of the enclosure on heat and
mass transfer will be discussed.

3. Numerical methods and validation

4.1. Vertical enclosure

A numerical solution procedure is adopted, which is depending

4.1.1. Influence of the double diffusive parameters
upon finite volume method to discretize the system of steady
Fig. 3 displays steady-state contour maps for various values of
equations on a staggered grid system. The third-order deferred cor-
RaT for N = 1, Da = 10
3, and Krp = Krs = 1.0. The condition N = 1.0
rection QUICK scheme and second-order central linear Interpola-
means that the fluid structure is dominated by equal and aiding
tion are, respectively, implemented for the convection and
effects of both thermal and solutal buoyancies. It is clearly
diffusion term. The resulting discretized equations are solved by
observed that a flow circulation in the clockwise direction is con-
a line-by-line procedure, combining the tri-diagonal matrix algo-
fined in the fluid and porous layer. As Ra is low (Ra = 103, Fig. 4
rithm (TDMA) and the successive over-relaxation (SOR) iteration.
(a)), the isotherms and iso-concentrations are almost vertical and
The pressure-velocity link is established through the SIMPLE algo-
parallel indicating that conduction is dominant. The fluid flow
rithm [47].
intensity in porous layer is relatively weak and the maximum tem-
In the present formulation, the computational domain was
perature is located in the center of porous layer, which shows that
divided into control volumes such that the control volume faces
thermal discharging of porous layer is slow. With the increase of
exactly coincide with the boundaries of the porous wall. The equa-
RaT, shown in Fig. 3(b) and (c), respectively, for Rat = 105, Rat =
tions are solved in the fluid and the porous layers simultaneously
with suitable modifications. In the porous-solid composite wall
the dimensionless diffusion coefficient in the energy equation is
changed from unity to Krp, Krs, respectively, while that in the mass
equation is changed from unity to Kc. By combing the energy equa-
tion in this manner, the matching conditions at fluid-porous and
porous-solid interfaces stated by Eqs. (29) and (30) are satisfied
automatically. The algorithm ensures continuity of the fluxes
across all control surfaces and, thus, the fluid-porous interfaces.
Convergence is ensured when maximal residual of all the govern-
ing equations become <10
5. In addition to the usual accuracy con-
trol, the accuracy of computation is also controlled using the
energy and species conservations within the system.
The validation of the numerical code was performed over a
large range of parameters. Typical results are presented in Table 1
for the case of enclosure partially filled with porous layer. In gen-
eral, the results are in good agreement with those of Singh et al. [4],
Baytas et al. [11] and Hadidi et al. [15], thereby providing valida-
tion to our simulations. Especially, the applied numerical method
for double diffusive natural convection of porous media has been
compared with the results of Singh et al. [4] for RaT = 105, N = 10,
Da = 10
3 and Le = 0.84 in Fig. 2, where the isotherms, streamlines (a)
and iso-concentrations demonstrate the accuracy of the utilized
code. A second validation was made through comparing present
results with Jiang-tao Hu [27] in the case of enclosure with a solid
partition. The good agreement is observed between the present
results with the published data in Table 2.
Grids ranged from (80  80) to (200  200) are tested and pre-
sented in Table 3 after some preliminary tests of asymptotic type
were carried out, and then the final grid resolution of (120 
120) was selected at the balance between the calculation accuracy
and the speed. Generally, each CFD simulation took about 2–3 h to
complete on Working Station (two Intel (R) Xeon (R) CPU E5 2690
processors, 256 G RAM).

4. Results and discussion

In the present study, the enclosure is square, i.e. AR = 1.0, and

fluid filled enclosures is the moist air, with a low concentration
of water vapor, it can be taken Pr = 0.71, Le = 0.8. Additionally,
the mass diffusion coefficient ratio is kept constant as Kcp = Kcs =
(b)
1.0. Attention is thus focused on the effects of the thermal Rayleigh
number Ra, buoyancy ratio N, Darcy number Da, thermal conduc- Fig. 7. Effect of Darcy number and buoyancy ratio on heat and mass transfer for Rat
tivity ratio Krp, Krs and angle of inclination of the enclosure U. = 106 and Krp = Krs = 1.0, (a) Nusselt number, (b) Sherwood number.
608 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
106, the flow becomes more intense with a vertical thermal and
solutal stratifications in the central area of the enclosure in Fig. 3
(b) and (c). By passing to higher values of the thermal Rayleigh
number (Fig. 3(d)), the fluid flow tends to be convection domi-
nated, where a primary flow eddy forms in fluid and porous layers.
Particularly, as thermal Rayleigh number discretely increases from
103 to 107, shown in Fig. 3(a)-(d), the fluid flow intensity is
strengthened the maximum temperature is also descending owing
to the enhanced convective effect.
To demonstrate the effect of the buoyancy ratio N on the heat
and moisture transport structures, thermal Rayleigh number and
Darcy number are fixed at 106 and 10
3, respectively. For N = 0.0
in Fig. 4(a), i.e. pure thermal-driven flow, the fluid circulates as a
large anti-clockwise rotating cell, moving upwards at the center
and downwards near the cold side walls. This is caused by volu-
metric heating that drives up the fluid in the central zone of the
enclosure while the symmetric cooling. The isotherms and iso-
concentrations are inclined and almost asymptotically adjacent
to the left cooling wall, and this is indicating that conduction is
dominant. Sparse streamlines in porous layer also show that con-
vection is relatively weak. As N decreases, shown in Fig. 4(b)-(c),
convective effect is more pronounced with dense streamlines and
(a) Krp = 0.1
(b) Krp = 0.5
(c) Krp = 5.0
(d) Krp = 10.0
Streamlines
Fig. 8. Contour lines of stream function, temperature, and concentration for Rat = 106, N = 1.0, Da = 10
3, Krs = 1.0, and U = 0°, (a) hmax = 0.003; (b) hmax = 0.013; (c) hmax =
0.128; (d) hmax = 0.2634.
significant concentration stratification in fluid and porous layers.
It is clearly observed that the buoyancy ratio contributes to the
great enhancement of heat and moisture transportation in this
enclosure with remarkably decreased maximum temperature.
However, when the buoyancy ratio is continually increasing, as
illustrated in Fig. 3(c) and (d), the flow structure, temperature and
concentration fields are center-symmetrical to those in Fig. 4(b)-
(c), respectively. Particularly, the maximum temperatures of these
situations are approximately equal to each other. This means that
thermal buoyancy forces always cooperate with each other in this
enclosure and the minus sign of solutal buoyancy ratio plays a vital
role in the fluid orientations.
The effect of RaT and N on the Nusselt and Sherwood numbers
are illustrated in Fig. 5 for Da = 10
3, and Krp = Krs = 1.0. When
RaT is small enough, the convection flow is much weak such that
the heat and mass transport in diffusion modes whatever the
buoyancy ratio is. As RaT increases, Nusselt and Sherwood numbers
are monotonously increasing and their minimum values are
located at N = 0, i.e. pure thermal-driven flow dominated. It is
observed that minimum values of RaT = 103, 104 and 105 both in
Nusselt and Sherwood number curves are almost identical due to
the fact that flow intensity is relatively weak in porous layer and
Isotherms Iso-concentrations
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
conduction is dominant in these situations. The positive impact RaT
is more pronounced as the absolute value of N increases in Fig. 5.
Moreover, the Nusselt and Sherwood numbers are increasing with
increasing the absolute value of buoyancy ratio and are symmetric
for N = 0, which also illuminate the augmenting effect of buoyancy
ratio in this enclosure.
As the Darcy number is low, typically 10
7 as shown in Fig. 6(a),
the boundary frictional resistance becomes progressively signifi-
cant and adds to the bulk frictional drag induced by the solid
matrix to slow the convection motion. This is expected because
in the limit of Da approaching 0, the Brinkman model reduces to
Darcy’s law. As illustrated in Fig. 6(a), a flow circulation in the
clockwise direction is confined in the fluid layer and the fluid flow
inside the porous media is negligible. This is due to the fact that
porous layer behaves like a solid wall and inhibits the heat and
moisture transports. When Darcy number increases to 10
5, the
fluid flow, temperature and concentration structures are similar
to the situation of Da = 10
7, which displays that the inhibition
effect of porous layer is negligible for Da < 10
5 (low permeability
region). As Darcy number step increases from 10
4 to 10
1 (Figs. 6
(c)-(d) and 3(c)), the effect of viscous forces accounted for in the
Brinkman term on the flow velocity becomes significant. As the
permeability of the porous medium Da increases, the boundary
(a) Krs = 0.1
(b) Krs = 0.5
(c) Krs = 5.0
(d) Krs = 10.0
Streamlines
Fig. 9. Contour lines of stream function, temperature, and concentration for Rat = 106, N = 1.0, Da = 10
3, Krp = 1.0, and U = 0°, (a) hmax = 0.0332; (b) hmax = 0.0286; (c) hmax =
0.0188; (d) hmax = 0.0173.
609
frictional resistance becomes gradually less important and the fluid
circulation within the porous layer is progressively enhanced.
Moreover, the maximum temperature shown in isotherms is
greatly reduced and the location of Tmax shifts toward the fluid
region, which indicating that the heat generated in porous materi-
als is dissipated well with increase of Darcy number.
Average heat and mass transfer rates as functions of buoyancy
ratio are presented in Fig. 7, with Darcy number as a parameter.
Heat and mass transfer rates are enhanced greatly with the buoy-
ancy ratio. In the low permeability range (Da < 10
5), Nusselt and
Sherwood numbers increase little, which is also concluded through
comparison of Fig. 6(a) and (b). However, as Da increases, the vari-
ation of Nu and Sh are essentially different. Sh always increases,
while Nu increases firstly in the solutal-driven flow region (N <
1.0 and N > 1.0) and then decreases in the thermal-driven flow
region (
1.0  N  1.0). That is, blindly increasing Da number of
this porous matrix will not always enhance heat and mass transfer
performance, which will guide the design of the thermal discharg-
ing progress.
4.1.2. Effects of thermal conductivity ratios Krp and Krs
The effect of porous and solid layers thermal diffusion
coefficients on the fluid, heat and moisture transport structures
Isotherms Iso-concentrations
610 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
are illustrated in Figs. 8 and 9, respectively, with that Da = 10
3,
RaT = 106 and N = 1.0 are maintained. As fluid-porous thermal con-
ductivity ratio is equal to 0.1, observed in Fig. 8(a), the fluid flow
intensity tends to be convection dominated, where the fluid rotates
as a large clockwise roll and vertical temperature and concentra-
tion stratifications are presented in the fluid layer. With the
increase of Krp, shown in Fig. 8(b) and (c), the flow becomes more
intense and the maximum temperature in isotherms decreases.
This means that enhancement of Krp could strengthen the fluid
flow, heat and moisture transports. However, further increasing
Krp to 10.0 in Fig. 8(d), heat generated inside the porous layer is
well removed and the total porous layer has a similar temperature,
show in isotherms, which leads to the competition of thermal and
solutal buoyancy forces. This causes that a clockwise eddy flow and
counterclockwise rotating recirculation zone respectively exist in
the central and upper section of this enclosure. It is also seen that
the maximum temperature is reduced and vertical concentration
stratification is destroyed.
(a)
(b)
Fig. 10. Effect of porous-fluid thermal conductivity ratio and buoyancy ratio on
heat and mass transfer for Rat = 106, Da = 10
3 and Krs = 1.0, (a) Nusselt number, (b)
Sherwood number.
Similarly, the thermal conductivity of solid layer has a positive
influence on the flow patterns and temperature filed structures,
observed in Fig. 9. As Krs is relatively low in Fig. 9(a), the highest
value of temperature shown in isotherms is mainly located in the
porous region and a jump exits in the porous/solid interface in iso-
therms due to the higher thermal conductivity of porous layer. The
location Tmax has a tendency to shift towards the fluid layer, when
Krs is up to 10.0 in Fig. 9(d), and temperature of total solid layer is
equal to 0.002, indicating thermal energy is well transferred. Fur-
thermore, the concentrations field structure varies little with the
modulation of Krs.
The evolution of heat and moisture transfer across the interface
between fluid and porous layers versus the buoyancy ratio at dif-
ferent values of Krp and Krs are presented in Figs. 10 and 11. Con-
cerning the Nusselt number, it may be observed that the increase
in heat transfer only occurs at lower Krp in this enclosure, while
the high value of Krs has a great influence. When Krp is <5.0,
displayed in Fig. 10(a), Nu is enhanced with increasing thermal
(a)
(b)
Fig. 11. Effect of solid-fluid thermal conductivity ratio and buoyancy ratio on heat
and mass transfer for Rat = 106, Da = 10
3 and Krp = 1.0, (a) Nusselt number, (b)
Sherwood number.
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
(a) Φ=15°
(b) Φ=45°
(c) Φ=75°
(d) Φ=135°
Streamlines
Fig. 12. Contour lines of stream function (top), temperature (center) and concentration (bottom) for Rat = 106, Da = 10
3, N = 1.0 and Krf = Krs = 1.0, (a) hmax = 0.023; (b) hmax =
0.0213; (c) hmax = 0. 0227; (d) hmax = 0. 0212.
conductivity of porous layer. Then increasing Krp up to 10.0, the
minimum of Nusselt number is shift towards N = 1.0 due to the fact
that there is a competition between thermal and solutal buoyancy
forces when N > 0. Consequently, the Nusselt number is firstly
boosted with increasing Krp and then it is decaying in this enclo-
sure. In other words, the peak value of Nusselt number is in the
vicinity of Krp = 5.0. However, as illustrated in Fig. 11(a), an
increase of Krs results in monotonously increasing, due to the
enhancement of heat transportation. Observing the Sh curves in
Figs. 10 and 11(b), both the thermal conductivities of porous and
solid layers have very little influence on moisture transfer rate.
Generally, an increase of thermal diffusive coefficient of porous
and solid layers will enhance heat removal from the energy storage
unit in some extent.
4.2. Inclined enclosure
The following part illustrates the effect of the angle of inclina-
tions U (0°  U  180°). Fig. 12 shows the contour maps for various
angle of inclinations for Rat = 106, Da = 10
3, N = 1.0 and Krf = Krs =
1.0. From Fig. 12(a) and (b), it is clearly observed that the flow filed
611
Isotherms Iso-concentrations
consists of a series of clockwise roll cells. Compared to the base
case with U = 0° (Fig. 3(c)), flow intensity increases considerably
and the maximum temperature decreases, making convection
more rigorous when the cooling wall is inclined. When U is
increased up to 75° shown in Fig. 12(c), flow intensity is still
enhanced, while the maximum temperature is declined. This is
expected because the direction of gravity is similar to those of ther-
mal and solutal buoyancy forces, and hence heat and moisture
transfer are inhibited. As U increases further, shown in Fig. 12(d),
a counter-clockwise eddy flow exists in the fluid and porous
regions, completely opposite to these situations of U < 90°. It is
interesting to note that temperature and concentration fields for
U = 135° shown in are symmetrical to those for U = 45°. The max-
imum temperature of these situations U = 135° and 45°, shown by
isotherms, are similar to each other.
Fig. 13 presents the results of Nusselt and Sherwood number as
a function of inclination angle for various values of thermal Ray-
leigh number. As U increases from 0° to 180°, Nu and Sh numbers
change a little for Rat < 105 due to the conduction dominant effect.
In convection dominant situations, i.e. Rat > 105, the maximum of
Nusselt number occurs at U = 45° and 135°, and the minimums
612 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
(a)
(b)
Fig. 13. Effect of thermal Rayleigh number and inclination angle on heat and mass
transfer for N = 1.0, Da = 10
3 and Krp = Krs = 1.0, (a) Nusselt number, (b) Sherwood
number.
of Nu and Sh are both located near U = 90°, indicating that the ori-
entation angle of U has an influence on heat and moisture trans-
port only when convection is dominant. It is also noticed from
Fig. 13 that the increase of thermal Rayleigh number leads to an
increase of both heat and moisture transfer rate, which indicates
that thermal Rayleigh number has a significant effect on Nu and Sh.
The variation of Nu and Sh with different inclinations U and
buoyancy ratios N is shown in Fig. 14. For N < 1.0, the Nusselt
and Sherwood number decrease firstly with the increase of U, pass-
ing through a minimum, and then begin to increase. The minimum
in Nusselt and Sherwood number occurs at 90°. In contrast, when N
is equal to 10.0, Nu curve in the range 0° < U < 90° increases firstly
and then decreases, and Sh curve monotonously increases and
achieves the peak at U = 90°. It is also observed that Nu and Sh
numbers of N = 10.0 are larger than those of N =
10.0. This fact
is illuminating that gravity and positive solutal force cooperate
each other in the inclined enclosure.
(a)
(b)
Fig. 14. Effect of buoyancy ratio and inclination angle on heat and mass transfer for
Rat = 106, Da = 10
3 and Krp = Krs = 1.0, (a) Nusselt number, (b) Sherwood number.
The combined effects of Darcy number Da and the angle of incli-
nation U on the Nusselt and Sherwood numbers is exhibited in
Fig. 15 for Rat = 106, N = 1.0 and Krp = Krs = 1.0. It is clearly seen
in Fig. 15(b) that Sh changes similarly compared with that shown
in Fig. 6 as the U increases. However, shown in Fig. 15(a), Nu curve
is presented as a sinusoid, where the peaks are located at U = 30°
and 150° and the trough is near at U = 90°, respectively. It is also
observed that Darcy number enhancing the heat transfer rate only
occurs in the range of 60° < U < 105°, which is useful in the appli-
cation of the inclined thermal storage wall system to determining
the optimized Darcy number.
The effect of thermal conductivity ratios Krf and Krs on Nu and
Sh are presented in Figs. 16 and 17 with Rat = 106, N = 1.0 and Da
= 10
3. As U increases, seen from Fig. 16(a), Nusselt number with
Krf = 10.0 increases firstly and then decreases, where the maximum
is located at U = 90°. When Krf < 10.0, Nu gets the maximum at U =
45° and 135°, respectively, while the minimum at U = 90°. This is
essentially attributed to the gravity decomposition. This situation
is similar in Fig. 17(a). As U is ranged from 0° to 90°, heat transfer
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
(a)
(b)
Fig. 15. Effect of Darcy number and inclination angle on heat and mass transfer for
Rat = 106, N = 1.0 and Krp = Krs = 1.0, (a) Nusselt number, (b) Sherwood number.
rate is enhanced firstly, and then inhibited. It is also noticed that
Nu decreases monotonically with the variation of Krf, while
increases monotonically with Krs, which are similar compared with
that shown in Figs. 10(a) and 11(a). Furthermore, an increase in Krf
and Krs results in little increasing of Sherwood number, which is
concluded in Figs. 10(b) and 11(b). As observed from Figs. 16(b)
and 17(b), for Krf  5.0 or Krs  5.0, Sh can increase abruptly; Krf
> 5.0 or Krs > 5.0, moisture transfer rate seems to be increasing lit-
tle with inclination angle.
5. Conclusions
Concerning on the thermal storage composite wall, double dif-
fusive convection partially filled with a porous layer with spatially
uniform internal heat generation and a solid layer in a square
enclosure has been investigated numerically. The thickness of por-
ous and solid layers are kept constant as 0.25. Fluid Prandtl num-
ber and Lewis number are set as 0.71 and 0.8, respectively. Fluid
flow, heat and moisture transport structures at a wide range of
613
(a)
(b)
Fig. 16. Effect of porous-fluid thermal conductivity ratio and inclination angle on
heat and mass transfer for Rat = 106, Da = 10
3 and Krs = 1.0, (a) Nusselt number, (b)
Sherwood number.
key parameters as Darcy number (Da) from 10
9 to 10
1, thermal
Rayleigh number (Rat) from 103 to 107, buoyancy ratio, N from

10.0 to 10.0, thermal conductivity ratio, Krf and Krs from 0.1 to
10.0, and inclination of the enclosure 0° < U < 90° have been
obtained and illustrated.
Under the situation of vertical enclosure, average Nusselt and
Sherwood numbers are an increasing function of Rat and N. The
thermal and solutal buoyancy forces are always aiding with each
other and the minus sign of N only has influence on the fluid orien-
tations. In the convective dominant regime, average Nusselt num-
ber increases in the solutal-driven flow region (N < 1.0 and N > 1.0)
and decreases in the thermal-driven flow region (
1.0  N  1.0)
with Darcy number increasing, while Sherwood number increases
monotonously. For lower permeability, heat and moisture transfer
rates do not vary substantially. Furthermore, an increase of Krp and
Krs result in the enhancement of average heat transfer rate, while
the thermal conductivities of porous and solid layers have very lit-
tle influence on moisture transfer rate. With the consideration of
angle of inclination, the average heat transfer rate as functions of
inclination angle, when N > 0.0, is presented like a sinusoid, where
the maximum is near at U = 45° and 135°and the minimums is
located at U = 90°. Generally, as inclination angle increases, mois-
ture transfer rate is decayed, and then enhanced, which gets the
minimum at U = 90°.
614 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
(a)
(b)
Fig. 17. Effect of solid-fluid thermal conductivity ratio and inclination angle on heat
and mass transfer for Rat = 106, Da = 10
3 and Krp = 1.0, (a) Nusselt number, (b)
Sherwood number.
Several simulations are made, varying the governing dimen-
sionless parameters and also the inclination angle of the enclo-
sures. A new and useful way is proposed for the presentation of
the average heat and moisture transfer rates. These results are use-
ful to understand the phenomena of heat and moisture transfer in
building structures with the thermal storage composite wall,
which could provide heating engineers with an effective design
tool against heat loss.
Conflict of interests
We declare that there is no conflict of interests regarding the
publication of this paper.
Acknowledgements
This research was financially supported by the Natural Science
Foundation of China (NSFC Grant No. 51778504; NSFC Grant No.
51208192; NSFC Grant No. 51304233), Hunan Provincial Natural
Science Foundation for Distinguished Young Scholars supported
by Hunan Provincial Government (Grant No. 14JJ1002), the Funda-
mental Research Funds for the Central Universities (Grant No.
2042016kf0136, Wuhan University) and the National Key Basic
Research Program from the Ministry of Science and Technology
of P. R. China (973 Program Grant No. 2014CB239203).
Both Prof. Fu-Yun Zhao and Prof. Han-Qing Wang would also
like to acknowledge the financial support from the Collaborative
Innovation Center for Building Energy Conservation and Environ-
ment Control, Hunan Province, China. Miss. Jiang-Tao Hu would
also like to acknowledge the support of China Scholarship Council.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.ijheatmasstransfer.
2018.02.071.
References
[1] Nasiru I. Ibrahim, Fahad A. Al-Sulaiman, Saidur Rahman, Bekir S. Yilbas, Ahmet
Z. Sahin, Heat transfer enhancement of phase change materials for thermal
energy storage applications: a critical review, Renew. Sustain. Energy Rev. 74
(2017) 26–50.
[2] Krishna Kota, Louis Chow, Quinn Leland, Laminar film condensation driven
latent thermal energy storage in rectangular containers, Int. J. Heat Mass
Transfer 55 (2012) 2308–11217.
[3] Souvik Pal, Muhammad R. Hajj, William P. Wong, Ishwar K. Puri, Thermal
energy storage in porous materials with adsorption and desorption of
moisture, Int. J. Heat Mass Transfer 69 (2014) 285–292.
[4] A.K. Singh, T. Paul, G.R. Thorpe, Natural convection due to heat and mass
transfer in a composite system, Heat Mass Transfer 35 (1999) 39–48.
[5] Mohamed Mharzi, Michel Daguenet, Sa.ad. Daoudi, Thermosolutal natural
convection in a vertically layered fluid-porous medium heated from the side,
Energ. Conv. Manage. 41 (2000) 1065–1090.
[6] Pinghua Zhao, C.F. Chen, Stability analysis of double-diffusive convection in
superposed fluid and porous layers using a one-equation model, Int. J. Heat
Mass Transfer 44 (2001) 4625–4633.
[7] F. Chen, C.F. Chen, Onset of finger convection in a horizontal porous layer
underlying a fluid layer, J. Heat Transfer 110 (1988) 403–409.
[8] R. Bennacer, H. Beji, A.A. Mohamad, Double diffusive convection in a vertical
enclosure inserted with two saturated porous layers confining a fluid layer, Int.
J. Therm. Sci. 42 (2003) 141–151.
[9] D. Gobin, B. Goyeau, J.-P. Songbe, Double diffusive natural convection in a
composite fluid-porous layer, J. Heat Mass Transfer 120 (1998) 234–242.
[10] D. Gobin, B. Goyeau, A. Neculae, Convective heat and solute transfer in
partially porous cavities, Int. J. Heat Mass Transfer 48 (2005) 1898–1908.
[11] A.C. Baytas, A.F. Baytas, D.B. Ingham, I. Pop, Double diffusive natural
convection in an enclosure filled with a step type porous layer: non-Darcy
flow, Int. J. Therm. Sci. 48 (2009) 665–673.
[12] K. Al-Farhany, A. Turan, Non-Darcy effects on conjugate double-diffusive
natural convection in a variable porous layer sandwiched by finite thickness
walls, Int. J. Heat Mass Transfer 54 (2011) 2868–2879.
[13] N. Hadidi, Y. Ould-Amer, R. Bennacer, Bi-layered and inclined porous collector:
optimum heat and mass transfer, Energy 51 (2013) 422–430.
[14] N. Hadidi, R. Bennacer, Y. Ould-Amer, Two-dimensional thermosolutal natural
convective heat and mass transfer in a bi-layered and inclined porous
enclosure, Energy 93 (2015) 2582–2592.
[15] N. Hadidi, R. Bennacer, Three-dimensional double diffusive natural convection
across a cubical enclosure partially filled by vertical porous layer, Int. J. Therm.
Sci. 101 (2016) 143–157.
[16] M. Ismael, F. Selimefendigil, A.J. Chamkha, Mixed convection in a vertically
layered fluid-porous medium enclosure with two inner rotating cylinders, J.
Porous Media 20 (2017) 491–511.
[17] C.Y. Han, S.W. Baek, The effect of radiation on natural convection in a
rectangular enclosure divided by two partitions, Numer. Heat Transfer, Part A
37 (2000) 249–270.
[18] F.Y. Zhao, D. Liu, G.F. Tang, Conjugate heat transfer in square enclosures, Heat
Mass Transfer 43 (2007) 907–922.
[19] C.D. Sankhavara, H.J. Shukla, Influence of partition length on natural
convection in partially divided square enclosures, J. Heat Transfer 129
(2007) 1592–1599.
[20] D. Liu, F.Y. Zhao, H.Q. Wang, Passive heat and moisture removal from a natural
vented enclosure with a massive wall, Energy 36 (2011) 2867–2882.
[21] V.A.F. Costa, Natural convection in partially divided square enclosures: effects
of thermal boundary conditions and thermal conductivity of the partitions, Int.
J. Heat Mass Transfer 55 (2012) 7812–7822.
[22] Hosni Souheil Harzallah, Abdessalem Jbara, Khalifa Slimi, Double-diffusive
natural convection in anisotropic porous medium bounded by finite thickness
 J.-T. Hu et al. / International Journal of Heat and Mass Transfer 123 (2018) 600–615
walls: validity of local thermal equilibrium assumption, Transp. Porous Med.
103 (2014) 207–231.
[23] A.J. Chamkha, M.A. Mansour, S.E. Ahmad, Double-diffusive natural convection
in inclined finned triangular porous enclosures in the presence of heat
generation/absorption effects, Heat Mass Transfer 46 (2010) 757–768.
[24] A.J. Chamkha, M.A. Ismael, Conjugate heat transfer in a porous cavity heated
by triangular thick wall, Numer. Heat Transfer, Part A 63 (2013) 144–158.
[25] A.J. Chamkha, M.A. Ismael, Conjugate heat transfer in a porous cavity filled
with nanofluids and heated by a triangular thick wall, Int. J. Therm. Sci. 67
(2013) 135–151.
[26] M. Sathiyamoorthy, A.J. Chamkha, Analysis of natural convection in a square
cavity with a thin partition for linearly heated side walls, Int. J. Numer.
Methods Heat Fluid Flow 24 (2014) 1057–1072.
[27] Hu. Jiang-Tao, Xiu-Hong Ren, Di Liu, Fu-Yun Zhao, Han-Qing Wang, Conjugate
natural convection inside a vertical enclosure with solid obstacles of unique
volume and multiple morphologies, Int. J. Heat Mass Transfer 95 (2016) 1096–
1114.
[28] Hu. Jiang-Tao, Xiu-Hong Ren, Di Liu, Fu-Yun Zhao, Han-Qing Wang, Natural
convective heat and moisture transfer in an inclined building enclosure with
one slender wall of finite thickness: analytical investigation and non-unique
steady flow solutions, Int. J. Heat Mass Transfer 104 (2017) 1160–1176.
[29] A. Ben-Nakhi, A.J. Chamkha, Natural convection in inclined partitioned
enclosures, Heat Mass Transfer 42 (2006) 311–321.
[30] R. Chowdhury, S. Parvin, M.A.H. Khan, A.J. Chamkha, MHD natural convection
in a porous equilateral triangular enclosure with a heated square body in the
presence of heat generation, Special Top. Rev. Porous Media 6 (2015) 353–365.
[31] K. Khanafer, A.J. Chamkha, Hydromagnetic natural convection from an inclined
porous square enclosure with heat generation, Numer. Heat Transfer, Part A 33
(1998) 891–910.
[32] A.J. Chamkha, Double-diffusive convection in a porous enclosure with
cooperating temperature and concentration gradients and heat generation or
absorption effects, Numer. Heat Transfer, Part A 41 (2002) 65–87.
[33] A.J. Chamkha, Hydromagnetic combined convection flow in a vertical lid-
driven cavity enclosure with internal heat generation or absorption, Numer.
Heat Transfer, Part A 41 (2002) 529–546.
[34] M.A. Mansour, A.J. Chamkha, R.A. Mohamed, M.M. Abd El-Aziz, S.E. Ahmed,
MHD natural convection in an inclined cavity filled with a fluid saturated
porous medium with heat source in the solid phase, Nonlinear Anal.: Model.
Control 15 (2010) 55–70.
[35] G.H.R. Kefayati, Simulation of double diffusive natural convection and entropy
generation of power-law fluids in an inclined porous cavity with Soret and
615
Dufour effects (Part I: Study of fluid flow, heat and mass transfer), Int. J. Heat
Mass Transfer 94 (2016) 539–581.
[36] G.H.R. Kefayati, Simulation of double diffusive natural convection and entropy
generation of power-law fluids in an inclined porous cavity with Soret and
Dufour effects (Part II: Entropy generation), Int. J. Heat Mass Transfer 94
(2016) 582–624.
[37] M.A. Ismael, A.J. Chamkha, Conjugate natural convection in a differentially
heated composite enclosure filled with a nanofluid, J. Porous Media 18 (2015)
699–716.
[38] M.A. Mansour, S. Elsayed, A.J. Chamkha, Entropy generation optimization for
MHD natural convection of a nanofluid in porous media-filled enclosure with
active parts and viscous dissipation, Int. J. Numer. Methods Heat Fluid Flow 27
(2017) 379–399.
[39] S.K. Jena, L.K. Malla, S.K. Mahapatra, A.J. Chamkha, Transient buoyancy-
opposed double diffusive convection of micropolar fluids in a square
enclosure, Int. J. Heat Mass Transfer 81 (2014) 681–694.
[40] K.M. Khanafer, A.J. Chamkha, Mixed convection flow in a lid-driven enclosure
filled with a fluid-saturated porous medium, Int. J. Heat Mass Transfer 42
(1999) 2465–2481.
[41] A.J. Chamkha, H. Al-Naser, Double-diffusive convection in an inclined porous
enclosure with opposing temperature and concentration gradients, Int. J.
Therm. Sci. 40 (2001) 227–244.
[42] A.J. Chamkha, H. Al-Naser, Hydromagnetic double-diffusive convection in a
rectangular enclosure with opposing temperature and concentration
gradients, Int. J. Heat Mass Transfer 44 (2002) 2465–2483.
[43] A.J. Chamkha, H. Al-Naser, Hydromagnetic double-diffusive convection in a
rectangular enclosure with uniform side heat and mass fluxes and opposing
temperature and concentration gradients, Int. J. Therm. Sci. 41 (2002) 936–
948.
[44] A.J. Chamkha, F. Abdulgafoor, Double-diffusive convection in a tilted enclosure
filled with a non-darcian porous medium, Int. J. Heat Technol. 24 (2006) 141–
152.
[45] M. Sathiyamoorthy, A.J. Chamkha, Effect of magnetic field on natural
convection flow in a square cavity for linearly heated side wall(s), Int. J.
Therm. Sci. 49 (2010) 1856–1865.
[46] A.J. Chamkha, A. Al-Mudhaf, E. Al-Meshaiei, Thermo-solutal convection in an
inclined porous cavity with various aspect ratios under mixed thermal and
species boundary conditions, Heat Transfer-Asian Res. 40 (2011) 693–720.
[47] S.V. Patankar, Numerical Heat Transfer and Fluid Flow, Hemisphere,
Washington DC, 1980.
[48] A. Bejan, Convection in Porous Media, fourth ed., Springer, 2013.