International Communications in Heat and Mass Transfer 53 (2014) 34–38

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

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

Surface radiation effect on convection in a closed enclosure driven by a

discrete heater☆
S. Saravanan ⁎, C. Sivaraj
Department of Mathematics, Bharathiar University, Coimbatore, Tamil Nadu, India

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

Available online 25 February 2014 This paper is aimed at presenting the changes experienced by a convective flow in a closed square enclosure
when surface radiation is taken into account. The flow is driven by a centrally placed discrete heater in an air filled
Keywords: two dimensional square enclosure. Symmetrically cooled isothermal vertical walls and insulated horizontal walls
Thermal radiation are considered. The governing coupled partial differential equations were solved using a finite volume method on
Natural convection a uniformly staggered grid system. The resulting augmentation of fluid velocities and the factors causing them are
Discrete heater
discussed.
Finite volume method
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction and radiation heat transfer in complex enclosures [4–7]. An overview of

Natural convection in rectangular enclosures is an area of study for
many researchers working on heat transfer in the past few decades
due to its wide range of applications. In several practical situations,
the temperature of the enclosure walls is significantly high so that the
contribution from surface radiation cannot be neglected in comparison
with natural convection. Although many experimental and numerical
studies have been carried in this area, more information is still needed
on the temperature distribution when the geometry involves an adia-
batic boundary. The influence of surface radiation is more significant
when the medium is filled with the transparent fluids. Surface to surface
radiation through a transparent medium modifies the temperature dis-
tribution of the adiabatic boundaries which in turn affects the tempera-
ture stratification in the enclosure and thereby the convective heat
transfer. Considerable works have already been performed on the
coupled natural convection and surface radiation in simple enclosures
[1–3]. The results obtained by these authors show that surface radiation
alters significantly the temperature distribution inside the enclosure
and the flow patterns. It is also found that the overall heat transfer
rate across the cavity is a strong function of emissivity of the enclosure
walls.
The interest has now shifted to complex enclosures with obstruc-
tions in the form of solid bodies or fins or partial baffles as their location,
length and thickness can significantly control the resulting flow charac-
teristics. Only few studies have been reported on the coupled convection
the above studies clearly shows that there are many other physical
setups of industrial interest for which the information on the effect of ra-
diation on convection is still lacking. The objective of the present study is
to discuss the effect of surface radiation on natural convective flow in-
duced by a discrete heater in a closed enclosure with cold vertical
walls. This type of basic configuration is often encountered in electronics
industry, nuclear and chemical energy production systems and solar en-
ergy collection systems.
2. Mathematical formulation
The physical model considered here is a two dimensional square en-
closure of length L containing a square discrete heater of length L/3 at its
center (see Fig. 1). The discrete heater is assumed to be isothermal at a
higher temperature Th. The vertical walls are cooled at a constant tem-
perature Tc while the horizontal walls are insulated. Air acts as the work-
ing medium within the enclosure. Non-emitting and non-absorbing
characters of air under moderate temperatures make it radiatively
non-participating. All enclosure walls and heater surfaces are assumed
to be opaque, gray and diffuse emitters and reflectors of radiation. The
flow is assumed to be laminar and incompressible. Then the governing
equations under the Boussinesq approximation are
∂u ∂v
þ ¼0 ð1Þ
∂x ∂y

☆ Communicated by W.J. Minkowycz. !

⁎ Corresponding author at: Department of Mathematics, Bharathiar University, 2 2
∂u ∂u ∂u 1 ∂p ∂ u ∂ u
Coimbatore 641 046, Tamil Nadu, India. þu þv ¼− þν þ ð2Þ
∂t ∂x ∂y ρ ∂x ∂x2 ∂y2
E-mail address: sshravan@lycos.com (S. Saravanan).

http://dx.doi.org/10.1016/j.icheatmasstransfer.2014.02.002
0735-1933/© 2014 Elsevier Ltd. All rights reserved.
 S. Saravanan, C. Sivaraj / International Communications in Heat and Mass Transfer 53 (2014) 34–38 35

Nomenclature y
Fkj view factor from the kth element to the jth element
g gravitational acceleration, ms−2 Insulated
Ik dimensionless irradiation of the kth element
thermal conductivity of fluid, W m−1 K−1 L
k
L length of the cavity, m
v
N total number of radiative surfaces g
NRC Radiation–conduction number (σT4hL/kΔT) u
p pressure, Pa
P dimensionless pressure (pL2/ρα2)
Pr Prandtl number (ν/α) Tc L Th Tc
q flux, W m−2 3
Q dimensionless flux
Ra Rayleigh number (gβΔTL3/αν)
Rk dimensionless radiosity of the kth element
t time, s L
T temperature, K
3
T0 average temperature ((Th + Tc)/2), K
u, v velocity components, m s−1 x
U, V dimensionless velocity components (uL/α, vL/α)
x, y cartesian coordinates, m
0 Insulated L
X, Y dimensionless coordinates (x/L, y/L)
Fig. 1. Physical configuration.

Introducing the following dimensionless variables

Greek symbols
α thermal diffusivity of fluid, m2 s−1
x y u v p t
β thermal expansion coefficient of fluid, K−1 X¼ ; Y¼ ; U¼ ; V¼ ; P¼ ; τ¼ 2
L L α=L α=L ρα 2 =L2 L =α
ΔT characteristic temperature difference, K ð6Þ
q T−T c
ε emissivity of the radiative surface Q rd ¼ rd4 ; θ ¼ ; where ΔT ¼ T h −T c
σ Th ΔT
Θ dimensionless temperature ratio (Tk/Th)
θ dimensionless temperature ((T − Tc)/ΔT)
the governing Eqs. (1)–(4) can be written in a dimensionless form as
ν kinematic viscosity of fluid, m2 s−1
ρ density of fluid, kg m−3
σ Stefan–Boltzmann constant, W K−4 m−2 ∂U ∂V
þ ¼0 ð7Þ
τ dimensionless time (αt/L2) ∂X ∂Y
Ψ dimensionless stream function (ψ/α)
ψ stream function, m2 s−1 !
∂U ∂U ∂U ∂P ∂2 U ∂2 U
þU þV ¼− þ Pr þ ð8Þ
∂τ ∂X ∂Y ∂X ∂X 2 ∂Y 2
Subscripts
c cold wall
h hot wall !
i incoming radiation ∂V ∂V ∂V ∂P ∂2 V ∂2 V
þU þV ¼− þ Pr þ þ RaPrθ ð9Þ
o outgoing radiation ∂τ ∂X ∂Y ∂Y ∂X 2 ∂Y 2
rd radiation

∂θ ∂θ ∂θ ∂2 θ ∂2 θ
þU þV ¼ þ ð10Þ
! ∂τ ∂X ∂Y ∂X 2 ∂Y 2
2 2
∂v ∂v ∂v 1 ∂p ∂ v ∂ v
þu þv ¼− þν þ þ gβðT−T c Þ ð3Þ
∂t ∂x ∂y ρ ∂y ∂x2 ∂y2 with the corresponding boundary conditions

! τ¼0: U ¼ V ¼ 0; θ ¼ 0; at 0 ≤X ≤1 and 0≤Y ≤1

∂T ∂T ∂T
2
∂ T ∂ T
2 τ N0 : U ¼ V ¼ 0; θ ¼ 0; at X ¼ 0; 1 and 0≤ Y ≤1
þu þv ¼α þ ð4Þ ∂θ
∂t ∂x ∂y ∂x2 ∂y2 U ¼ V ¼ 0; ¼ NRC Q rd ; at Y ¼ 0 and 0bXb1
∂Y :
∂θ
with the boundary conditions U ¼ V ¼ 0; ¼ −NRC Q rd ; at Y ¼ 1 and 0bXb1
∂Y
U ¼ V ¼ 0; θ ¼ 1; on the heater
t¼0: u ¼ v ¼ 0; T ¼ Tc; at 0≤x ≤L and 0 ≤y≤L ð11Þ
t N0 : u ¼ v ¼ 0; T ¼ Tc; at x ¼ 0; L and 0≤y ≤L
∂T qrd The dimensionless parameters appearing in the Eqs. (7)–(11) are
u ¼ v ¼ 0; ¼ ; at y ¼ 0 and 0bxbL
∂y k : ð5Þ defined as
∂T −qrd
u ¼ v ¼ 0; ¼ ; at y ¼ L and 0bxbL
∂y k ν gβΔTL
3 4
σ ThL
u ¼ v ¼ 0; T ¼ Th; on the heater Pr ¼ Ra ¼ NRC ¼ : ð12Þ
α να k
36 S. Saravanan, C. Sivaraj / International Communications in Heat and Mass Transfer 53 (2014) 34–38

In order to obtain the fluid flow in terms of streamlines, the dimen- 4. Discussion of results
sionless stream function Ψ is calculated through
The present study is made to analyze the interaction between sur-
∂Ψ ∂Ψ face radiation and buoyancy induced flow in a square enclosure con-
U¼ ; V ¼− : ð13Þ taining a discrete heater. Computations were carried out for Pr = 0.72
∂Y ∂X
corresponding to air. The characteristic temperature difference ΔT and
the average temperature of air were chosen to be 20 K and 303.15 K re-
For the combined convection and radiation problem in an enclosure
spectively. The effects of Ra and ε are investigated. All the surfaces are
containing a transparent medium, the radiative heat transfer can arise
considered to have the same emissivity εk = ε which can be taken as ε
only via thermal boundary conditions. In the present problem, the ther-
= 0 representing the absence of radiative heat exchange and ε = 1
mal boundary conditions at the insulated walls include the contribution
representing the perfect absorption and emission of radiative flux, i.e.,
of the net radiative flux. In order to determine the radiative flux, the net-
black surface. We focus on the steady state results alone and plot the
radiation method [8] is used. Thus, the walls of the enclosure and the
isotherms and streamlines side by side exploiting the symmetry about
discrete heater are divided into N number of radiative surface elements
the vertical center line.
such that these surface elements coincide with the control volume
It is seen from Eq. (11) that the surface radiation is coupled with con-
solid–fluid interfaces. Therefore, the net radiative flux of the kth ele-
vection only through the top and bottom boundary conditions. In partic-
ment is expressed as
ular the convective/radiative flux entering an insulated surface is
balanced by the radiative/convective flux of equal magnitude leaving
qrd;k ¼ qo;k −qi;k ; k ¼ 1…N ð14Þ
the surface. It is obvious that in the present configuration under investi-
gation the inclusion of surface radiation creates identical and uniform
where the subscripts o and i refer respectively to the outgoing and convex temperature profiles instantaneously, with their maxima at
incoming components. They are defined as their centers, at the top and bottom insulated walls. However the pres-
ence of convection within the enclosure has a tendency to affect the uni-
XN
4
form convex profiles and makes them nonuniform as the steady state is
qo;k −ð1−ε k Þ F kj qo; j ¼ εk σT k ; k ¼ 1…N ð15Þ approached. One may refer to our earlier study to have an idea on the
j¼1
transient temperature distribution [7]. In order to understand heat en-
ergy exchange near the insulated walls at the steady state we consider
the net radiative fluxes Qrd at the top and bottom walls (Fig. 2). They ex-
X
N
hibit almost similar parabolic patterns when convection is very weak
qi;k ¼ F kj qo; j ; k ¼ 1…N ð16Þ
(Ra = 103). In general Qrd is positive in the middle region and negative
j¼1
elsewhere. It shows that the outward radiative flux is stronger than the

where εk is the emissivity of the kth element and Fkj is the view factor
from the kth element to the jth element of the radiative surface
elements. Using the definitions Qrd,k = qrd,k/σT4h, Rk = qo,k/σT4h, Ik =
q i,k /σT 4h and Θ k = T k /T h, we obtain the dimensionless form of net
Ra 10 3 Bottom wall

radiative flux, radiosity and irradiation respectively: Ra 10 7 Top wall

0.04
Q rd;k ¼ Rk −Ik ; k ¼ 1…N ð17Þ

XN
4
Rk −ð1−ε k Þ F kj R j ¼ εk Θk ; k ¼ 1…N ð18Þ
j¼1

0.00
X
N
Ik ¼ F kj R j ; k ¼ 1…N: ð19Þ Qrd
j¼1

3. Solution methodology

The governing Eqs. (7)–(10) were discretized by the finite volume -0.04
method on a uniform staggered grid system using the SIMPLE algo-
rithm. The third order QUICK scheme [9] and the second order central
difference scheme were, respectively, used for the convection and diffu-
sion terms. In order to keep consistent accuracy over the entire compu-
tational domain, a third order accurate boundary condition treatment
was adopted as suggested in [9]. The set of discretized equations were
then solved by a line-by-line procedure of the tri-diagonal matrix algo-
rithm (TDMA). For radiation calculations, the net-radiation method was -0.08
0.0 0.2 0.4 0.6 0.8 1.0
used. Then the corresponding system of radiosity Eq. (18) was solved by X
the Gauss elimination method. The view factors were determined by
using Hottel's crossed-string method [10]. For the code validation and Fig. 2. Variation of dimensionless radiative flux along the horizontal walls for ε = 1 and
others one may refer our earlier paper [7]. different Ra.
 S. Saravanan, C. Sivaraj / International Communications in Heat and Mass Transfer 53 (2014) 34–38 37

inward one in the middle region and the later dominates the former in
the outer regions adjoining the sidewalls.
At large Ra (Ra = 107) the net radiative fluxes lose their parabolic
pattern and become distinct. Throughout the bottom wall Qrd is nega-
tive implying that it receives additional heat energy through the inward
radiative flux. Hence the resulting hotter bottom wall generates convec-
tive flux. This additional heating of the bottommost fluid layer through
natural convection enhances the buoyancy mechanism already present
in the enclosure in the absence of radiation. In contrary Qrd remains pos-
itive almost throughout the top wall, except near the sidewalls. One may
expect this because of the enhanced convective flux hitting the top wall
at such a high Ra. One should notice that it also includes the additional
fluid flux generated at the bottom wall due to the radiative heating.
This results in energy loss of the fluid layer adjacent to the top wall pro-
ducing a cooling down effect on it which was not observed in the ab-
sence of radiation. Thus the incorporation of surface radiation on the
enclosure imparts additional heating and additional cooling effects on
the fluid layer adjacent to the bottom and top walls respectively.
We also notice from Fig. 2 that Qrd is decreasing as one moves away
from the center indicating a reduction in the convective flow adjoining
the top wall and thus an increase in the cooling effect. The relevant iso-
therms and streamlines are shown in Fig. 3 for two different values of Ra
and ε. As anticipated they resemble each other to a great extent for
lower Ra. One can notice the characteristic changes introduced by the
effects discussed earlier on the thermal and streamline distributions.
The isothermal pattern below the discrete heater indicates the signifi-
cantly improved buoyancy. The additional heating has caused increased
fluid fluxes near the bottom wall which in turn makes the lower half of
the cells stronger. As a result a stronger fluid flux hits the top wall which
in turn increases the radiative outflux from there causing a substantial
reduction in its temperature. One should note that this is the reason
for the appearance of slanting isotherms near the top wall. Thus in the
presence of radiation the increased fluid flux, after hitting the top
wall, moves horizontally to greater distances when compared to the
case ε = 0. One should notice that the topmost fluid stream loses its en-
ergy gradually and hence descends near the cold wall which in turn
makes the primary eddy near the top wall elongated. Thus convective
outflux acts in regions very near the sidewalls which can be observed
through Fig. 2 as well. However one should note that these two effects,
viz., additional radiative heating and additional convective cooling show
their presence only for large Ra.
5. Conclusion
The combined influence of surface radiation and natural convection
on the fluid flow inside a square enclosure containing a discrete heater
at its center has been investigated numerically. Surface radiation shows
its effect only for higher values of Ra. In this situation the net radiative
fluxes at the top and bottom walls become quite distinct. The bottom
wall receives energy through the incoming radiative flux and generates
additional buoyancy which in turn makes the bottom portions of the
convective cells stronger. The additional fluid flux approaching the top
wall affects its heat content. The top wall loses its energy through radi-
ation and makes the isotherms inclined inwards which in turn causes

-0.01 -0.01
0.37

0.42

-0 . 8 -0.
4 09 0.6 09
0.7

Ra = 10 3

0.68 -2
5.0 63
8 0.
-30.56

1
0. 2
0.05
-1.4 7
-1. 80

7
Ra = 10
(a) ε = 0 (b) ε = 1
Fig. 3. Isotherms and streamlines for different Ra and ε.
38 S. Saravanan, C. Sivaraj / International Communications in Heat and Mass Transfer 53 (2014) 34–38
stronger convective motion adjoining the top insulated wall, stretching
the primary eddies up to the cold wall. These two changes homogenize
the temperature field throughout the enclosure and increase the local
fluid velocities.
Acknowledgment
The authors thank the University Grants Commission, India for its
support through the DRS Special Assistance Programme in Fluid Dy-
namics. One of the authors (C.S.) thanks the CSIR, India for its financial
support through the Senior Research Fellowship.
References
[1] M. Akiyama, Q.P. Chong, Numerical analysis of natural convection with surface radi-
ation in a square enclosure, Numer. Heat Transfer, Part A 31 (1997) 419–433.
[2] C. Balaji, S.P. Venkateshan, Interaction of surface radiation with free convection in a
square cavity, Int. J. Heat Fluid Flow 14 (3) (1993) 260–267.
[3] N. Ramesh, S.P. Venkateshan, Effect of surface radiation on natural convection in a
square enclosure, J. Thermophys. Heat Transf. 13 (3) (1999) 299–301.
[4] Y. Liu, N. Phan-Thien, A complete conjugate conduction convection and radiation
problem for a heated block in a vertical differentially heated square enclosure,
Comput. Mech. 24 (1999) 175–186.
[5] A. Mezrhab, H. Bouali, H. Amaoui, M. Bouzidi, Computation of combined
natural-convection and radiation heat-transfer in a cavity having a square
body at its center, Appl. Energy 83 (2006) 1004–1023.
[6] H. Sun, E. Chénier, G. Lauriat, Effect of surface radiation on the breakdown of steady
natural convection flows in a square, air-filled cavity containing a centered inner
body, Appl. Therm. Eng. 31 (2011) 1252–1262.
[7] S. Saravanan, C. Sivaraj, Coupled thermal radiation and natural convection heat trans-
fer in a cavity with a heated plate inside, Int. J. Heat Fluid Flow 40 (2013) 54–64.
[8] R. Siegel, J.R. Howell, Thermal Radiation Heat Transfer, fourth ed. Taylor and Francis
Group, New York, 2002.
[9] T. Hayase, J.A.C. Humphrey, R. Greif, A consistently formulated QUICK scheme for
fast and stable convergence using finite-volume iterative calculation procedures, J.
Comput. Phys. 98 (1992) 108–118.
[10] H.C. Hottel, A.F. Saroffim, Radiative Heat Transfer, McGraw Hill, New York, 1980.