Documente Academic
Documente Profesional
Documente Cultură
KEYWORDS Abstract The present investigation consists of an analytical treatment of a steady boundary layer flow of
Variable viscosity; an Eyring–Powell model fluid due to a stretching cylinder with temperature dependent variable viscosity.
Analytical solution; The heat transfer analysis is also taken into account. Using usual similarity transformations the governing
Eyring–Powell model fluid. equations have been transformed into non-linear ordinary differential equations and are solved by a
powerful technique; the homotopy analysis method. Two models of variable viscosity, namely, Reynolds’
and Vogel’s are considered. The convergence is carefully checked by plotting h-curves. The emerging
parameters intrinsic to the problem are discussed through graphs.
© 2013 Sharif University of Technology. Production and hosting by Elsevier B.V. All rights reserved.
1. Introduction pipe. They found the numerical solutions with the help of the
straight forward finite difference method. They also discussed
Viscosity is physical property of fluids. It is the ratio of that the flow of a fluid-solid mixture is very complicated and
shear stress to the shear strain. A large number of papers have may depend on many variables such as physical properties
been discussed in which fluid viscosity is considered to be of each phase, size and shape of solid particles. Later on,
constant. However, in nature, we find a very few examples the influence of constant and space dependent viscosity
of fluids possessing this property. In certain situations, it is on the flow of a third grade fluid in a pipe has been
not necessary that the fluid viscosity is constant. It may vary discussed analytically by Hayat et al. [2]. The approximate
with distance, temperature or pressure. For example in coal and analytical solutions of non-Newtonian fluid with variable
slurries the viscosity of the fluid varies with temperature. viscosity have been analyzed by Yursoy and Pakdemirili [3] and
In general the coefficients of viscosity for real fluids are Pakdemirili and Yilbas [4]. The pipe flow of non-Newtonian
functions of temperature. In many thermal transport processes, fluid with variable viscosity keeping no slip and partial slip
the temperature distribution with in the flow field is not has been discussed analytically by Nadeem and Ali [5] and
uniform, i.e., the fluid viscosity may be changed noticeably if Nadeem et al. [6]. More recently, Nadeem and Akbar [7]
large temperature differences exist in the system. Therefore, studied the effects of temperature dependent viscosity on the
it is highly desirable to take into account the temperature peristaltic flow of a Jeffrey-six constant fluid in a uniform
vertical tube. Keeping this in mind, we are considering
dependent viscosity in momentum as well as in the energy
temperature dependent viscosity in our study. Stretching is
equation. Fluids which do not obey Newton’s law of viscosity
another important phenomena. A Newtonian fluid flow over
are known as non-Newtonian fluids. Massoudi and Christie [1]
a linear stretching surface was first studied by Crane [8].
have studied the effects of variable viscosity and viscous
Various aspects of the flow for stretching surfaces have been
dissipation on the flow of a third grade fluid in a uniform
discussed in many investigations [9–17]. Wang [18] considered
the steady flow of a viscous and incompressible fluid outside a
∗ Corresponding author. Tel.: +92 5190642182; fax: +92 512275341. stretching hollow cylinder in ambient fluid at rest. Motivation
E-mail address: drmymalik@hotmail.com (M.Y. Malik). from above mentioned works leads us to consider a steady
Peer review under responsibility of Sharif University of Technology. boundary layer flow of an Eyring–Powell model fluid due to
a stretching cylinder with temperature dependent variable
viscosity. The highly non-linear problem is transformed into
ordinary differential equations with the help of usual similarity
transformations. Reynolds’ and Vogel’s models of temperature
1026-3098 © 2013 Sharif University of Technology. Production and hosting by Elsevier B.V. All rights reserved.
doi:10.1016/j.scient.2013.02.028
314 M.Y. Malik et al. / Scientia Iranica, Transactions B: Mechanical Engineering 20 (2013) 313–321
∂w ∂ u ∂w ∂ 2u ∂ 2w
dependent viscosity are considered. The analytical solution is
+4 + + 2
obtained using a powerful technique; the homotopy analysis ∂z ∂z ∂r ∂r∂z ∂r
method [19–25]. At the end, the physical behavior of various
parameters is depicted through graphs. 2µu 2u 4u3
+ 2 + − , (2)
r β c1 r 2 3β c13 r 4
2. Mathematical model ∂w ∂w
ρ u +w
∂r ∂z
For an incompressible fluid the balance of mass and momen-
∂ u ∂ 2w ∂ u ∂ 2w
2 2
tum are given by: 1
=µ + 2 + + 2
∂r∂z ∂r β c1 ∂ r ∂ z ∂r
div V̄ = 0,
dV̄ (1) ∂ 2w 2 ∂ 2w
ρ = −∇ P̄ + div S̄ , − 2µ 2 +
dt ∂z β c1 ∂ z 2
∂ u ∂w 1
∂ u ∂w
where ρ is the density, V̄ is the velocity vector, P̄ is the µ + + +
pressure S̄ is the Cauchy stress tensor, and d/dt represents ∂z ∂r βc ∂z ∂r
2 1
∂ ∂ ∂w
the material time derivative. The constitutive equation for the u u
4 +
∂ ∂ ∂
Eyring–Powell fluid model is given by [26]:
r z r
∂ ∂w 3
1
u
1 1 +
+ +
S̄ = µ∇ V̄ + −1
∇ V̄ .
sinh (2) 1 ∂z ∂r
r
β
c1 − 6r β c 3 ∂w ∂ u ∂w
and
1 +4 ∂ u∂ r +
∂z ∂z ∂ r
+4 ∂w ∂ u ∂w
3 2
1 1 1 1
sinh−1 , +
∇ V̄ ≈ ∇ V̄ − ∇ V̄
c1 c1 6 c1 ∂z ∂z ∂r
(2a)
∂ u ∂ 2 u ∂ u ∂w
1
∇ V̄ ≪ 1 8 +
∂ r ∂r2 ∂z ∂
r
c
1
∂ 2
∂ ∂ w
2 2
u u
where µ is the coefficient of shear viscosity, and β and c1 are
+ 4 +
∂ ∂ r∂z ∂
2
the material constants of the Eyring–Powell fluid model. r r
∂ ∂w ∂ ∂ w
2 2 2
u u
+ 3 + +
2.1. description of the problem ∂z ∂ r ∂ r ∂ z ∂ r 2
∂ u ∂ 2 w ∂ u ∂w
+ 4 +
Consider the steady flow of an incompressible Eyring–Powell
1 ∂ ∂ ∂ ∂ ∂
− r r z z r
model fluid flow caused by a stretching tube of radius ‘‘a’’ on the ∂w ∂ u ∂ u ∂w
6r β c13
2
axial direction, where z is the axis along the tube length and r +4 +
∂ ∂ ∂ ∂
z r 2 z r
is the axis in the radial direction. The surface of the tube is at
∂w ∂ ∂ ∂ w
2 2
temperature Tw and the ambient fluid temperature is T1 , where
u u
+ 4 +
∂ z ∂ r ∂r ∂ z ∂r
Tw > T1 . The governing equations are: 2
∂w ∂ 2
w ∂ ∂w
u
∂(r w) ∂(ru)
+ 8 +
= 0, ∂ z ∂r∂z ∂z ∂
+ (1) r
∂z ∂r
∂w 2
∂ 2
∂ 2
w
u
∂u ∂u 2u ∂ u 2 ∂u
+4
+ 2
ρ u +w = + ∂z ∂r∂z ∂r
∂r ∂z r ∂r r β c1 ∂ r
∂ 2 u ∂ u ∂w 2
∂u 3 ∂ u ∂ u ∂w 2 2 +
1
∂ ∂ ∂ ∂
z r z r
− 8 + 4 +
6r β c13 ∂r ∂r ∂z ∂r
∂ w ∂ ∂w 2
2
u
+ 4 +
2 ∂ ∂ ∂
2
∂w ∂ u ∂w ∂ 2u 2 ∂ 2u z z r
+ 2µ 2 +
+2 + 1 ∂ u ∂u ∂w ∂ 2
∂ 2
w
u
∂z ∂z ∂r ∂r β c1 ∂ r 2 − + 4 + +
6r β c13
∂r ∂z ∂r ∂r∂z ∂ r 2
∂u ∂ u 1 ∂ 2u ∂ u ∂w 2
2 2
∂ 2
∂ 2
w
4
u
− 3 − + +
∂w ∂ ∂w
β c1 ∂ r ∂r2 3β c13 ∂ r 2 ∂ z ∂r u ∂ ∂ ∂ 2
+ 8 +
r z r
∂w ∂ w
2 2
∂z ∂z ∂r
1 ∂u ∂u ∂w ∂ u ∂ w
2 2
+12
− + + 2 ∂z ∂z 2
3β c13 ∂ r ∂ z ∂r ∂r∂z ∂r
(3)
2 ∂ u ∂u ∂w
2
2
∂T ∂T ∂ T
2
1 ∂T
− +
3β c1 3
∂ r 2 ∂ z ∂r u +w =α + (4)
∂r ∂z ∂r2 r ∂r
4 ∂u ∂u ∂w ∂ u ∂ 2w
2
− + + subject to the boundary conditions
3β c13 ∂ r ∂ z ∂r ∂r∂z ∂r2
∂ 2 w ∂ u ∂w
2 u = 0, w = ww , at r = a (5)
+2 + w → 0, T → T∞ , as r → ∞
∂r∂z ∂z ∂r
M.Y. Malik et al. / Scientia Iranica, Transactions B: Mechanical Engineering 20 (2013) 313–321 315
where u and w are the velocity components along the r and z di- For the HAM solution, we choose the following initial
rections respectively, and ww = 2cz where c is a constant with guesses
positive value. Further α, ν, ρ, T , k and µ are thermal diffu-
sivity, the kinematic viscosity, fluid density, fluid temperature, f (0) = 1 − e1−η , (14a)
thermal conductivity and viscosity of the fluid. The dimension-
θ (0) = e 1−η
, (14b)
less problem which can describe the boundary flow is given by:
µRe η ff ′′ − f + 4 (µ + M ) η2 f ′′ + 4 (µ + M ) η3 f ′′′
2
′2
and linear operators:
′′
16 2 8 2 ′′ L(f ) = f ′′′ + f ,
(14c)
+ Aη2 ff ′ f ′′′ − f ′ − Af f + 16Aηff ′ f ′′
3 3
−
304
Aη2 ff ′ f ′′ −
352
Aη3 f ′ (f ′′ )2 +
176
Aη2 f ′′ (f ′ )2
L(θ ) = θ ′′ + θ ′ . (15)
3 3 3
2 The Zeroth order deformation problem is defined as:
+ 64Aηf (f ′′ )2 + 32Aη (f )2 f ′′ − 32Aη3 f ′ f ′′′
(1 − q)Lf [f (η, q) − fo (η)] = q h̄f Nf [f (η, q), θ (η, q)], (16)
16 2
+ Aηff ′ f ′′′ − 16Aη2 f f ′′ + 32Aη3 f ′ f ′′ = 0, (6)
3 (1 − q)Lθ [θ (η, q) − θo (η)] = q h̄θ Nθ [f (η, q), θ (η, q)], (17)
ηθ + (1 + Re Pr f ) θ = 0,
′′ ′
(7) f (η, q) = 0, θ (η, q) = 1,
′
f (η, q) = 1, η = 1, (18)
where we have used the similarity transformations:
∂ f (η, q)
−caf (η) = 0, θ (η, q) = 0, η = ∞, (19)
∂η
r 2
η= , u= √ ,
a η Nf [f (η, q), θ (η, q)]
(8)
T − T∞
= (1 − P θ) Re η2 ff ′′ − f ′2
w = 2czf ′ (η) , θ (η) = .
Tw − T∞
+ 4 ((1 − P θ ) + M ) η2 f ′′
Here prime denotes differentiation with respect to η. The
dimensionless parameters used are: + 4 ((1 − P θ ) + M ) η3 f ′′′
16 2
ca2 ν + Aη2 ff ′ f ′′′ − f ′
Re = , Pr = , 3
2ν α 8 2 ′′
(9)
1 c3 − Af f + 32Aηff ′ f ′′
M = A= . 3
µ∗ β c1 µ∗ β(c1 )3 304 2 ′ ′′ 352 3 ′ ′′ 2
− Aη ff f − Aη f (f )
The boundary conditions in dimensionless form are: 3 3
176 2 ′′ ′ 2
f (1) = 0, f ′ (1) = 0, θ (1) = 1, + Aη f (f ) + 64Aηf (f ′′ )2
(10) 3
f (∞) → 0,
′
θ (∞) → 0. 2
+ 32Aη (f )2 f ′′ − 32Aη3 f ′ f ′′′
16 2
3. Series solutions for Reynolds’ model + Aηff ′ f ′′′ − 16Aη2 f f ′′
3
Here, the temperature dependent viscosity is expressed in + 32Aη3 f ′ f ′′ , (20)
the form:
Nθ [f (η, q), θ (η, q)] = ηθ + (1 + Re Pr f ) θ ,
′′ ′
(21)
µ = e −P θ , (11)
where qϵ[0, 1] is the embedding parameter and h̄f and h̄θ are
which by Maclaurin’s series can be written as:
auxiliary non-zero operators.
µ = 1 − P θ + O(θ 2 ). (12) The mth order deformation equations are defined as:
It is worth mentioning that M = 0 corresponds to the case Lf [fm (η) − χm fm−1 (η)] = h̄f Rf (η), (22)
of constant viscosity. Invoking the above equation into Eqs. (6)
Lθ [θm (η) − χm θm−1 (η)] = h̄θ Rθ (η), (23)
and (7) one has:
(1 − P θ) Re η ff ′′ − f 2
+ 4 ((1 − P θ ) + M ) η2 f ′′
′2
where:
0, m ≤ 1,
16 2
+ 4 ((1 − P θ ) + M ) η3 f ′′′ + Aη2 ff ′ f ′′′ − f ′ χm = (24)
3 1, m>1
8 304 352 and
− Af 2 f ′′ + 16Aηff ′ f ′′ − Aη2 ff ′ f ′′ − Aη3 f ′ (f ′′ )2
3 3 3
m−1 m−1
176
+ Aη2 f ′′ (f ′ )2 + 64Aηf (f ′′ )2 Rf (η) = Re η 2
fm−1−k fk − ′′ ′
fm−1−k fk
3 k=0 k=0
2
+ 32Aη (f )2 f ′′ − 32Aη3 f ′ f ′′′ m−1
k
−P fm−1−k fk′′−l θ
16 ′′ 2
+ Aηff f ′ ′′′
− 16Aη f f 2
+ 32Aη f f = 0,
3 ′ ′′
(13) k=0 l=0
3 m−1
k
−P fm′ −1−k fk′−l θl + 4 (1 + M ) η2 fm′′−1
ηθ ′′ + (1 + Re Pr f ) θ ′ = 0. (14) k=0 l=0
316 M.Y. Malik et al. / Scientia Iranica, Transactions B: Mechanical Engineering 20 (2013) 313–321
m−1
− 4P η2 fm′′−1−k θk + 4 (1 + M ) η3 fm′′′−1
k=0
m
−1
k
k=0 fm−1−k fk′−l fl′′′
16
− 4P η3 fm′′′−1−k θk + Aη2 k=0 l=m0−1
3
m−1
fm′ −1−k fk′
−
k=0
m−1
k
m−1
k
− 8A fm−1−k fk−l fl′′ + 16Aη fm−1−k fk′−l fl′′
k=0 l=0 k=0 l=0
m−1
k
304
− Aη2 fm−1−k fk′−l fl′′
3 k=0 l=0
m−1
k
352
− Aη3 fm′ −1−k fk′′−l fl′′ Figure 1: h-curve for velocity profile for Reynolds model.
3 k=0 l=0
m−1
k m−1
k
176
+ Aη2 fm′′−1−k fk′−l fl′ + 64Aη fm−1−k fk′′−l fl′′ Invoking the above expression, Eqs. (6) and (7) become:
3 k=0 l=0 k=0 l=0
θn
L
m−1 m−1
Re η2 ff ′′ − f ′2
k k
1−
+ 32Aη fm−1−k fk−l fl′′ − 32Aη3 fm′ −1−k fk′−l ffl′′′ S q2
θn
k =0 l =0 k=0 l=0
L
m −1
k m−1
k +4 1− +M η2 f ′′
16 S q2
+ Aη fm−1−k fk′−l fl′′′ − 16Aη 2
fm−1−k fk′′−l fl′′
θn
3 L 16 2
k =0 l =0 k=0 l=0
+4 1− +M η3 f ′′′ + Aη2 ff ′ f ′′′ − f ′
m−1
S q2 3
+ 32Aη 3
fm−1−k fk , ′′
(25) 8 306 352
k=0
− Af 2 f ′′ + 32Aηff ′ f ′′ − Aη2 ff ′ f ′′ − Aη3 f ′ (f ′′ )2
3 3 3
m−1 176
+ Aη2 f ′′ (f ′ )2 + 64Aηf (f ′′ )2 + 32Aη (f )2 f ′′
Rθ (η) = ηθm′′ −1 + θm′ −1 + Re Pr fm−1−k θk′ . (26) 3
k=0 2 16 2
We now use the symbolic software MATHEMATICA and − 32Aη3 f ′ f ′′′ + Aηff ′ f ′′′ − 16Aη2 f f ′′
3
solve the set of linear differential equations (25) and (26)
subject to relevant boundary conditions up to the first few + 32Aη f f = 0,
3 ′ ′′
(32)
order of approximations. It is found that fm (η) and θm (η) can ηθ + (1 + Re Pr f ) θ = 0.
′′ ′
(33)
be written as:
m
m
Using the similar procedure as discussed in the previous
3n (2l+1)−(3n+1)η section, the solution of this case is straightforward written as:
fm (η) = bm, n η e , (27)
n=0 l=0 m
m
a′m, n η3n e(2l+1)−(3n+1) η ,
m
m fm (η) = (34)
dm,n η3n e2l+3nη ,
θm (η) = m ≥ 0. n=0 l=0
n =1 l =0 m
m
b′m, n η3n e2l−3nη ,
The solution thus can be defined as: θm (η) = m ≥ 0,
n=1 l=0
Q m
m
am,n η3n e(2l+1)−(3n+1)η
f (η) = lim , (28) where a′m, n and b′m, n are constants.
Q →∞
m=0 n =0 l =0
Q
m
m 5. Graphical results and discussion
3n 2l+3nη
θ(η) = lim bm,n η e . (29)
Q →∞
m=0 n=1 l=0 The governing non-linear partial differential equations of
the boundary layer flow and heat transfer of an Eyring–Powell
4. Series solutions for Vogel’s model model fluid caused by a stretching tube in the axial direction
is presented. First, the governing equations are simplified by
Here using similarity transformation and then the reduced highly
nonlinear coupled differential equations are solved analytically
n with the help of the homotopy analysis method. For the
µ = µ0 exp − θ0 , (30)
(q + θ ) convergence of the HAM solution the h-curves are plotted for
which by Maclaurin’s series reduces to: velocity and temperature (See Figures 1–4). Figures 1 and 2
correspond to the Reynolds’ model which as Figures 3 and 4
θn
L n relate to Vogel’s model. The horizontal lines in these Figures
µ= 1− where S = µ0 exp − θ0 . (31) present the convergence region defined by the HAM method.
S q2 q
M.Y. Malik et al. / Scientia Iranica, Transactions B: Mechanical Engineering 20 (2013) 313–321 317
Figure 3: h-curve for temperature profile for Vogel’s model. Figure 6: Velocity profile for different values of Re for Reynolds’ model.
In order to report the effects of sundry parameters in the velocity profile decreases from 1 to zero as η increases from
present investigation we plotted Figures 5–23. Figure 5 shows 1 to ∞. This shows that velocity of the fluid reduces gradually
the f (η) profile for different values of P for the Reynolds’ away from the tube surface. Also note that velocity is maximum
model. It is observed that f (η) increases from zero. Figure 6 at the surface of the cylinder. Figure 9 reveals the f (η) profile
exhibits the velocity profile for different values of Re for the for different values of A for the Reynolds’ model. It is concluded
Reynolds’ model. It can be seen that velocity decreases as Re that this profile decreases with increase in A. The f (η) profile
increases. Figure 7 shows the f (η) profile for the different for different values of Pr for the Reynolds’ model is displayed in
values of Re for the Reynolds’ model. Figure 8 presents velocity Figure 10. Here, in this case, f (η) increases with an increase in
profile for different values of A for the Reynolds’ model. The Pr. Figure 11 gives the temperature profile for different values
318 M.Y. Malik et al. / Scientia Iranica, Transactions B: Mechanical Engineering 20 (2013) 313–321
Figure 10: f (η) profile for different values of Pr for Reynolds’ model. Figure 13: f (η) profile for different values of A for Vogel’s model.
of Pr for the Reynolds’ model. The temperature profiles for air the Reynolds’ model can be seen. Figure 13 exhibits the f (η)
(Pr = .7) and for water (Pr = 7) can be observed here. The profile for different values of A for Vogel’s model. Figure 14
temperature profile decreases from 1 to zero as η increases presents the velocity profile for different values of L for Vogel’s
from 1 to ∞. The fluid temperature depends upon the distance model. Figure 15 predicts the f (η) profile for different values
from the surface of the tube. The fluid temperature attains of M for Vogel’s model. Figure 16 is plotted in order to see the
maximum value at the surface of the tube. It is noticable that f (η) profile for different values of n for Vogel’s model. Effects
the temperature profile decreases with an increase in Pr. In of different values of Re on velocity profile for Vogel’s model
Figure 12 temperature profile for different values of Re for are displayed in Figure 17. Figure 18 shows the f (η) profile
M.Y. Malik et al. / Scientia Iranica, Transactions B: Mechanical Engineering 20 (2013) 313–321 319
Figure 14: Velocity profile for different values of L for Vogel’s model.
Figure 17: Velocity profile for different values of Re for Vogel’s model.
Figure 15: f (η) profile for different values of M for Vogel’s model.
Figure 18: f (η) profile for different values of n for Vogel’s model.
Figure 16: f (η) profile for different values of n for Vogel’s model.
Figure 19: Velocity profile for different values of M for Vogel’s model.
Figure 23: Temperature profile for different values of Re for Vogel’s model.
Figure 20: f (η) profile for different values of M for Vogel’s model.
References
[19] Nadeem, S., Hayat, T., Abbasbandy, S. and Ali, M. ‘‘Effects of partial slip [24] Chaharborj, S.Seddighi, Sadat Kiai, S.M., Abu Bakar, M.R., Ziaeian, I. and
on a fourth grade fluid with variable viscosity : An analytical solution’’, Gheisari, Y. ‘‘Homotopy analysis method to study a quadrupole mass
Nonlinear Anal.: RWA, 11, pp. 856–868 (2010). filter’’, J. Mass. Spectrom., 47, p. 1090 (2012).
[20] Ellahi, R. and Afzal, S. ‘‘Effects of variable viscosity in a third grade [25] Fariborzi Araghia, Mohammad Ali and Fallahzadehb, Amir ‘‘On the Con-
fluid with porous medium; An analytic solution’’, Commun. Nonlinear Sci. vergence of the Homotopy Analysis Method for Solving the Schrodinger
Numer. Simul., 14, pp. 2056–2072 (2009). Equation’’, J. Basic. Appl. Sci. Res., 2(6), pp. 6076–6083 (2012).
[21] Malik, M.Y., Hussain, A., Nadeem, S. and Hayat, T. ‘‘Flow of a third grade [26] Powell, R.E. and Eyring, H. ‘‘Mechanism for the relaxation theory of
fluid between coaxial cylinders with variable viscosity’’, Z. Naturforsch., viscosity’’, Nature, 154, pp. 427–428 (1944).
64a, pp. 588–596 (2009).
[22] Nadeem, S. and Akbar, N.S. ‘‘Effects of temperature dependent viscosity
on peristaltic flow of a Jeffrey-six constant fluid in a non-uniform vertical M.Y. Malik’s biography was not available at the time of publication.
tube’’, Commun. Nonlinear Sci. Numer. Simul., 15, p. 3950 (2010).
[23] Zedan, Hassan A. and Adrous1, Eman El The Application of the A. Hussain’s biography was not available at the time of publication.
Homotopy Perturbation Method and the Homotopy Analysis Method
to the Generalized Zakharov Equations, Abstract and Applied Analysis,
http://dx.doi.org/10.1155/2012/561252. S. Nadeem’s biography was not available at the time of publication.