This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights
Author's personal copy
ORIGINAL ARTICLE
Thermodynamic optimization of uid ow over
an isothermal moving plate
A. Malvandi
a,
*
, F. Hedayati
b
, D.D. Ganji
b
a
Mechanical Engineering Department, Amirkabir University of Technology, Tehran, Iran
b
Department of Mechanical Engineering, Islamic Azad University, Sari Branch, Sari, Iran
Received 24 February 2013; revised 1 June 2013; accepted 20 June 2013
Available online 13 July 2013
KEYWORDS
Boundary layer;
Moving plate;
Second-law analysis;
Entropy generation minimi-
zation;
Similarity solution
Abstract In this paper, entropy generation minimization (EGM) was employed in order to achieve a
thermodynamic optimization of uid owand heat transfer over a at plate. The basic boundary layer
equations including continuity, momentum, energy, and entropy generation have been reduced to a
two-point boundary value problemvia similarity variables and solved numerically via RungeKutta
Fehlberg scheme. The novelty of this study was to consider the effects of velocity ratio k which rep-
resents the ratio of the wall velocity to the free stream uid velocity in a thermodynamic system.
Focusing on the velocity ratio as a pivotal parameter, in view of minimizing the entropy generation,
the optimum value of k k
o
was achieved. Moreover, considering Bejan number, it was shown that
the region, in which the maximum entropy generates, gets closer to the plate as k increases.
2013 Production and hosting by Elsevier B.V. on behalf of Faculty of Engineering, Alexandria
University.
1. Introduction
Optimal consumption of energy depends on designing an
efcient heat transfer process from thermodynamic point
of view. The main issue in this eld is to determine the
thermodynamically optimum size or operating regime of a
certain engineering system, where by optimum we mean
the condition in which the system losses the least power
while still performing its fundamental engineering function.
It turns out that in many systems, various mechanisms
and design features that account for irreversibility compete
with one another. Accordingly, thermodynamic optimization
that concerns us here is the condition of the most desirable
trade-off between two or more competing irreversibility.
Design methodology, known as entropy generation
minimization (EGM), has been comprehensively covered by
Bejan [1] specically, in the elds of refrigeration, heat trans-
fer, storage, solar thermal power conversion, and thermal
science education. The EGM is also known as second-law
analysis, thermodynamic optimization, and thermodynamic
design, or by new names such as exoirreversible, or endore-
versible thermodynamics [2]. This eld has experienced
astounding growth during the 1980s and 1990s, in both engi-
neering and physics. The popularity of EGM can be gauged
from several numbers of the researches done by scientists on
thermodynamic optimization in convective ows that
continue to appear in the heat transfer literature [37] .
*
Corresponding author. Tel.: +98 9370370330.
E-mail address: amirmalvandi@aut.ac.ir (A. Malvandi).
Peer review under responsibility of Faculty of Engineering, Alexandria
University.
Production and hosting by Elsevier
Alexandria Engineering Journal (2013) 52, 277283
Alexandria University
Alexandria Engineering Journal
www.elsevier.com/locate/aej
www.sciencedirect.com
1110-0168 2013 Production and hosting by Elsevier B.V. on behalf of Faculty of Engineering, Alexandria University.
http://dx.doi.org/10.1016/j.aej.2013.06.006
Author's personal copy
Nowadays, the classical boundary layer ow has been con-
sidered by many researchers in which the viscous forces are
important and may affect the engineering process of produc-
ing. Blasius [8] studied the simplest boundary layer over a at
plate by employing a similarity transformation to reduce the
partial differential boundary layer equations to a nonlinear
third-order ordinary differential equation. Extension of this
study has been considered by Sakiadis [9] who introduced
the boundary layer ow induced by a moving plate in a quies-
cent ambient uid. Later, Klemp and Acrivos [10], Hassanien
[11], Fang and Zhang [12] and recently Ishak et al. [13] and
Cortell [14] have put the focus on the ow over a moving plate.
A large amount of literature on this problem has been cited in
the books by Schlichting and Gersten [15] and White [16].
Thermal boundary layer ow over a moving surface was stud-
ied both analytically and experimentally by Tsou et al. [17].
Then, Lee and Davis [18] solved the general problem of com-
pressible boundary layer ow and temperature induced by a
moving continuous plane and axisymmetric surface numeri-
cally. Revankar [19] considered transient heat transfer from
a continuous moving plate with step change in variable wall
temperature. Next, many researchers developed this concept
through the years [2024]. It is worth mentioning that ows in-
duced by moving boundary with a parallel free stream have
many industrial applications such as heat treatment of material
traveling between a feed roll and windup roll or conveyor
belts, extrusion of steel, melt spinning process in the extrusion
of polymers continuous casting, glass blowing, cooling of a
large metallic plate in a bath [19,21,23].
Recently, second-law analysis of uid ow and heat
transfer across a at plate has been conducted by Malvandi
et al. [25]. They considered the effects of Prandtl number,
Eckert number, and Reynolds number simultaneously;
although some increasing/decreasing trends for governing
parameters were observed, they could not nd any optimum
case in which the entropy generation is minimized. To the
best of the authors knowledge, there is no investigation
on the thermodynamic optimization of the ow over a mov-
ing plate yet. In this paper, in order to minimize the gener-
ated entropy over a moving plate, we have employed the
similarity variables introduced by Blasius [8] and obtained
a relation for entropy generation inside the boundary layers.
It turns out that focusing on velocity of the plate, an opti-
mum system may be reached. Hence, we set velocity of the
plate as a pivotal parameter and studied its effects on entro-
py generation.
2. Governing equation
2.1. Velocity and temperature
Let us consider an incompressible viscous ow over a moving
at plate which moves in the same and also the opposite direc-
tions to the free stream with different velocities. The physical
model is presented in Fig. 1, and the detailed geometries have
been shown in Figs. 2ac. The wall temperature, T
w
, is uniform
and constant and is greater than the free stream temperature,
T
1
. It is assumed that the free stream velocity, U
1
, is also
uniform and constant. The governing equations including
Nomenclature
C
f
skin friction coefcient
f dimensionless stream function
c
p
specic heat at constant pressure
k thermal conductivity
Nu local Nusselt number
Pr Prandtl number
q
w
surface heat ux
Re local Reynolds number
T uid temperature
u velocity
S dimensionless entropy generation
S
000
gen
volumetric entropy generation rate
Ec Eckert number
Be Bejan number
Sh entropy generation due to heat transfer
S
f
entropy generation due to uid friction
Greek symbols
a thermal diffusivity
h dimensionless temperature
k velocity ratio
m kinematic viscosity
l dynamic viscosity
q uid density
s shear stress
w stream function
g similarity variable
u dissipation term in energy equation
Subscripts
w condition at the surface of the wedge
1 ambient conditions
o optimum value
c critical value
Figure 1 Physical ow model.
278 A. Malvandi et al.
Author's personal copy
continuity, momentum, and energy equations in the Cartesian
coordinates system can be expressed as:
@u
@x

@v
@y
0 1
u
@u
@x
v
@u
@y
t
@
2
u
@y
2
2
u
@T
@x
v
@T
@y
a
@
2
T
@y
2
3
Subject to the following boundary conditions:
At y 0 : v 0; u u
w
and T T
w
lim
y!1
u u
1
; lim
y!1
T T
1
4
We look for a similarity solution of the Eqs. (1) and (2) with
the boundary conditions of Eq. (4) in the following form:
w

2mxu
1
_
fg 5
g

u
1
2mx
_
y 6
where w is the usual stream function, i.e., u
@w
@y
and v
@w
@x
and m is the kinematic viscosity of the uid.
Substituting Eqs. (5) and (6) into Eq. (2), following ordin-
ary differential equation was obtained:
f
000
g fgf
00
g 0 7
With the boundary conditions:
f0 0; f
0
0 k
lim
g!1
f
0
g 1 8
where k
uw
u1
is the velocity ratio of the plate that represents the
ratio of the wall velocity to the free stream. More discussion on
the characteristics of k can be found in references [11,22].
The skin friction coefcient C
f
can be dened as:
C
f

s
w
qu
2
1
9
where s
w
is the surface shear stress which is given by:
s
w
l
@u
@y
_ _

y0
10
Substituting Eqs. (5), (6) into Eqs. (9) and (10) led to:

2Re
p
C
f
f
00
0 11
where Re
qu1x
l
is the local Reynolds number.Similarly, Eq.
(3) in similarity form can be expressed as:
h
00
g Prfgh
0
g 0 12
where
h
T T
1
T
w
T
1
13
is the dimensionless temperature and Pr
m
a
. Also, the bound-
ary conditions can be expressed as:
Figure 3 Dimensionless velocity inside the boundary layer for
different values of k.
Figure 2 Velocity boundary layer on a plate when: (a) k < 0:, (b)
0 < k < 1:, and (c) k > 1.
Thermodynamic optimization of uid ow over an isothermal moving plate 279
Author's personal copy
At g 0; h0 1 14
lim
g!1
hg h1 0
The local Nusselt number Nu
x
can be dened as:
Nu
xq
w
kT
w
T
1

15
where q
w
is the surface heat ux which can be read as:
q
w
k
@T
@y

y0
16
Using Eqs. (5), (6), (15) and (16), heat ux can be written as:
Re=2

1
2
Nu h
0
0 17
Like [22], in the present context, [Re/2]
1/2
Nu and

2Re
p
C
f
are referred as the reduced Nusselt number and reduced skin
friction coefcient which are represented by h
0
(0) and f
00
0,
respectively.
2.2. Entropy generation
Referring to [1], the entropy generation equation can be writ-
ten as
S
000
gen

k
T
2
rT
2
_ _

l
T
/ 18
Substituting Eqs. (5), (6), and (13) into Eq. (18), the entropy
generation equation for a ow over a moving plate can be ex-
pressed as:
S
h
0
g
2
2RePrEch
1
hg
2
g
2
2Re
1
_ _

1
2
f
00
g
2
h
1
hgRe
_ _
19
where the dimensionless entropy generation may be dened by
the following relationship:
S
m
2
DT
u
4
1
l
S
000
gen
20
In Eq. (19), the rst term is because of heat transfer and the
second one is due to uid friction which are called S
h
and S
f
,
respectively; Bejan number is dened as follows:
Be
S
h
S
h
S
f
21
As it is obvious, Be yields the share of S
h
and S
f
in total
generated entropy, Be = 1 is the limit at which the heat trans-
fer irreversibility dominates, Be = 0 is the opposite limit at
which the irreversibility is dominated by uid friction effects,
and Be = 0.5 is the case that the heat transfer and uid friction
entropy generation rates are equal. More physical interpreta-
tions of this concept can be found in [1].
3. Results and discussion
The system of Eqs. (7) and (12) with boundary condition equa-
tions (Eqs. (8) and (14)) has been solved numerically with Run-
geKuttaFehlberg method. Having considered concept of k,
the problem is studied for k < 0; k < 1 and k > 1 (Figs. 24).
In all calculations, here, we have assumed that h
1
= 2,
Re = 1000, Pr = 1, Ec = 0.01 with no loss of generality.
Rigorous readers can nd more information on the effects of
Prandtl, Eckert, and Reynolds numbers in [25].
Fig. 2a shows schematic velocity prole for a moving plate
in opposite direction to free stream velocity, i.e., k < 0. The re-
gion above the plate can be divided into two zones, namely
Zone 1 and Zone 2. Zone 1 thickens the boundary layer; as
the plate moves faster, Zone 1 gets thicker and solution stabil-
ity decreases. According to Hassanien [11], there exists a criti-
cal plate velocity ratio k
c
< 0 beyond which no solution exists.
Here, the critical velocity prole was obtained k
c
0:35410
which is in the best agreement with Hassanien [11]. Fig. 2b
shows the hydrodynamic boundary layer over a plate when it
moves slower than the free stream velocity, i.e., 0 < k < 1. In
this case, due to less difference between the minimum and max-
imum velocity vectors, boundary layer gets thinner. This trend
continues till the plate moves faster than the uid, i.e., k > 1
when plate exerts a pulling force to the uid next to it, so
the friction force get inverse and boundary layer thickness in-
creases with increasing in k; see Fig. 2c.
Fig. 3 shows how velocity prole of the uid can be affected
by the plates velocity. Also, it shows the accuracy of schematic
boundary layers demonstrated in Fig. 2. As it is obvious,
hydrodynamic boundary layer proles are different for k < 1
and k > 1. In the case of k > 1, uid velocity gradually ap-
proaches to free stream velocity in transverse direction (g).
However, for k < 1, uid velocity takes an increasing trend
with moving outward from the wall. In this case, in spite of
k > 1, due to low momentum transfer inside the boundary
layer, thickness of the boundary layer decreases.
The variation in velocity gradients with k has been shown in
Fig. 4. Clearly, F
00
has contrastive signs for k > 1 and k < 1.
When k > 1, the plate exerts a pulling force to the uid, and
hence, velocity gradients are negative. Here, as k increases,
absolute values of F00 increase as well. On the ip side for
k < 1, the velocity of the plate is less than that of the free
stream. Therefore, friction force grows between the plate and
the uid. In this case, when k < 0, a gentle rise can be seen
Figure 4 Dimensionless velocity gradients inside the boundary
layer for different values of k.
280 A. Malvandi et al.
Author's personal copy
in F
00
g which occurs at the inection point of the velocity dis-
tribution in Fig. 2a.
Reduced skin friction coefcient versus k is plotted in
Fig. 5. Obviously, since there is no shear stress at k 1, skin
friction coefcient vanishes C
f
= 0. Moreover, the values of
reduced skin friction coefcient depend on k; positive and neg-
ative values of C
f
correspond to k < 1 and k > 1, respectively.
Focusing on physical concept, the negative sign of the reduced
skin friction coefcient means that the plate exerts a drag force
on the uid and the positive sign implies the opposite. Needless
to say that at k 0, i.e., stationary plate, reduced skin friction
coefcient reaches its peak.
Variation in dimensionless temperature prole and its gra-
dients with k are displayed in Figs. 6 and 7, respectively. As
mentioned before, when k increases, momentum transfer near
the plate intensies which leads to stronger heat transfer and
thinner boundary layer; hence, temperature gradients inten-
sify. Not surprisingly, heat transfer coefcient takes an increas-
ing trend with k, which is fully supported by Fig. 8. As it was
expected, reduced Nusselt number varies more noticeably
while k < 1.
Values of Bejan number for different k inside the boundary
layer have been plotted in Fig. 9. Clearly, the values of Bejan
number are uctuating around the unity, which means that
heat transfer irreversibility dominates. Due to the equality be-
tween velocity of the plate and uid at k 1;, there is no
hydrodynamic boundary layer at all and consequently, heat
transfer between the plate and uid exists only as a result of
Figure 5 Reduced skin friction coefcient for different values of
k.
Figure 7 Dimensionless temperature gradients inside the bound-
ary layer for different values of k.
Figure 6 Dimensionless temperature inside the boundary layer
for different values of k.
Figure 8 Reduced Nusselt number for different values of k.
Thermodynamic optimization of uid ow over an isothermal moving plate 281
Author's personal copy
temperature gradient (conduction mechanism). Consequently,
only heat transfer irreversibility generates the total entropy.
Fig. 10 shows the values of generated entropy inside the
boundary layer. As it is obvious, after a gentle rise, the amount
of generated entropy decreases and nally drops to zero. As
stated before, near the wall, temperature and velocity gradients
are almost constant; however, the value of dimensionless tem-
perature has a decreasing trend. Based on Eq. (19), it is not
surprising that the amount of generated entropy inside the
boundary layer increase. It is worth mentioning that the higher
k gets, the distance between the critical point and the surface
increases, which widen the range of the constant velocity gra-
dient. Therefore, as k increases, the peak value of the generated
entropy curve gets closer to the surface.
Finally, total entropy generation versus k inside the bound-
ary layer has been presented in Fig. 11. It can be seen that an
optimum k
o
0:32 exists, in which the entropy generation is
minimized. Needless to say that as k increases, total generated
entropy inside the boundary layer climbs up as well. It is note-
worthy that at k 1, only heat transfer irreversibility gener-
ates the whole entropy.
4. Conclusion
This survey deals with minimization of entropy generation in-
side the boundary layer over a moving at plate. Mentioned
problem has many industrial applications such as heat treat-
ment of material traveling between a feed roll and windup roll,
extrusion of steel, glass blowing, and cooling of a large metallic
plate in a bath. The transformed ODE equations have been
solved with RungeKuttaFehlberg as a reliable numerical
technique. Focusing on the velocity ratio as a pivotal parame-
ter, its optimum value at which entropy generation is minimum
was achieved. Moreover, considering Bejan number, it was
shown that the region, at which the maximum entropy gener-
ates, gets closer to the plate when it moves faster.
References
[1] A. Bejan, Entropy Generation Minimization, second ed., CRC,
Boca Raton, Florida, 1996.
[2] K. Yakut, B. Sahin, S. Canbazoglu, Performance and ow-
induced vibration characteristics for conical-ring turbulators,
Applied Energy 79 (1) (2004) 11.
[3] S. Ahmet, A. Vedat, Entropy production in boundary layers,
Journal of Thermophysics and Heat Transfer 4 (3) (1990) 3.
[4] A. Bejan, Entropy Generation Through Heat and Fluid Flow,
Wiley, New York, 1982.
[5] W.A. Khan, R.S.R. Gorla, Entropy generation in non-
Newtonian uids along a horizontal plate in porous media,
Journal of Thermophysics and Heat Transfer 25 (2) (2011) 6.
[6] S.K. Rastogi, D. Poulikakos, Second-law optimization of forced
convection of non-Newtonian uids in ducts, Journal of
Thermophysics and Heat Transfer 6 (3) (1992) 3.
[7] G.F. Naterer, J.A. Camberos, Entropy and the second law uid
ow and heat transfer simulation, Journal of Thermophysics
and Heat Transfer 17 (3) (2003) 11.
Figure 10 Generated entropy inside the boundary layer for
different values of k.
Figure 11 Total entropy generation.
Figure 9 Bejan Number for different values of k.
282 A. Malvandi et al.
Author's personal copy
[8] H. Blasius, Grenzschichten in Flu ssigkeiten mit kleiner Reibung,
Z. Math. Phys. 56 (1908) 37.
[9] B.C. Sakiadis, Boundary-layer behavior on continuous solid
surfaces. I. Boundary-layer equations for two-dimensional and
axisymmetric ow, AIChE Journal 7 (1961) 2628.
[10] J.B. Klemp, A. Acrivos, A method for integrating the boundary-
layer equations through a region of reverse ow, Journal of
Fluid Mechanics 53 (1972) 14.
[11] I.A. Hassanien, Flow and heat transfer on a continuous at
surface moving in a parallel free stream of power-law uid,
Applied Mathematical Modelling 20 (10) (1996) 779784.
[12] T. Fang, J. Zhang, An exact analytical solution of the Falkner
Skan equation with mass transfer and wall stretching,
International Journal of Non-Linear Mechanics 43 (2008) 6.
[13] A. Ishak, R. Nazar, I. Pop, Flow and heat transfer
characteristics on a moving at plate in a parallel stream with
constant surface heat ux, International Journal of Heat Mass
Transfer 45 (2009) 4.
[14] R. Cortell, Flow and heat transfer in a moving uid over a
moving at surface, Theoretical in Computational and Fluid
Dynamics 21 (2007) 11.
[15] H. Schlichting, K. Gersten, Boundary Layer Theory, eigth ed.,
Springer, Berlin, 2000.
[16] F.M. White, Viscous FluidFlow, McGraw-Hill, NewYork, 2006.
[17] F.K. Tsou, E.M. Sparrow, R.J. Goldstein, Flowand heat transfer in
the boundary layer on a continuous moving surface, International
Journal of Heat and Mass Transfer 10 (2) (1967) 219235.
[18] W.W. Lee, R.T. Davis, Laminar boundary layers on moving
continuous surfaces, Chemical Engineering Science 27 (12)
(1972) 21292149.
[19] S.T. Revankar, Heat transfer to a continuous moving at
surface with variable wall temperature, International Journal of
Heat and Fluid Flow 10 (4) (1989) 357360.
[20] A.-S. Sami A, Mixed convection heat transfer along a
continuously moving heated vertical plate with suction or
injection, International Journal of Heat and Mass Transfer 47
(67) (2004) 14451465.
[21] M.-I. Char, C.O.-K. Chen, J.W. Cleaver, Conjugate forced
convection heat transfer from a continuous, moving at sheet,
International Journal of Heat and Fluid Flow 11 (3) (1990) 257
261.
[22] N. Bachok, A. Ishak, I. Pop, Boundary-layer ow of nanouids
over a moving surface in a owing uid, International Journal of
Thermal Sciences 49 (9) (2010) 16631668.
[23] T.C. Chiam, Magnetohydrodynamic boundary layer ow due to
a continuously moving at plate, Computers & Mathematics
with Applications 26 (4) (1993) 17.
[24] M. Martin, I. Boyd, Momentum and heat transfer in a laminar
boundary layer with slip ow, Journal of Thermophysics and
Heat Transfer 20 (4) (2006) 9.
[25] A. Malvandi, D.D. Ganji, F. Hedayati, M.H. Kaffash, M.
Jamshidi, Series solution of entropy generation toward an
isothermal at plate, Thermal Science 16 (5) (2012) 1289
1295.
Thermodynamic optimization of uid ow over an isothermal moving plate 283