You are on page 1of 10

INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

Flow and heat transfer over an unsteady
stretching sheet in a micropolar fluid with
prescribed surface heat flux
N. Bachok, A. Ishak, and R. Nazar

Abstract— The unsteady laminar flow of an incompressible
micropolar fluid over a stretching sheet with prescribed surface heat
flux is investigated. The governing partial differential boundary layer
equations are first transformed into ordinary differential equations
before being solved numerically by a finite-difference method. The
effects of the unsteadiness parameter, material parameter and Prandtl
number on the flow and heat transfer characteristics are studied. It is
found that the surface shear stress and the heat transfer rate at the
surface are higher for micropolar fluids compared to Newtonian
fluids.

Keywords—Unsteady flow, Heat transfer, Stretching sheet,
Micropolar fluid, Fluid mechanics.
I.

INTRODUCTION

T

HE fluid dynamics due to a stretching surface is important
in manufacturing processes. Examples are numerous and
they include the aerodynamic extrusion of plastic sheets, the
boundary layer along a liquid film in condensation processes,
paper production, glass blowing, metal spinning and drawing
plastic films. The thermal fluid flow problems have been
extensively studied numerically, theoretically as well as
experimentally (see [1,2]). The quality of the final product
depends on the rate of heat transfer at the stretching surface.
Crane [3] first obtained an elegant analytical solution to the
boundary layer equations for the problem of steady twodimensional flow due to a stretching surface in a quiescent
incompressible fluid. Gupta and Gupta [4] extended the
problem posed by Crane [3] to a permeable sheet and obtained
closed-form solution, while Grubka and Bobba [5] studied the
thermal field and presented the solutions in terms of Kummer’s
functions. The 3-dimensional case has been considered by
Wang [6]. Chen [7] studied the case when buoyancy force is
taken into consideration, and Magyari and Keller [8]
considered exponentially stretching surface. The heat transfer
over a stretching surface with variable surface heat flux has
Manuscript received July 28, 2010. This work was supported in part by the
IPTA Fundamental Research Grant.
N. Bachok is with the Department of Mathematics and Institute for
Mathematical Research, Universiti Putra Malaysia, 43400 UPM Serdang,
Selangor, Malaysia (phone: 603-8946-6849; fax: 603-8943-7958; e-mail:
norfifah@math.upm.edu.my).
A. Ishak and R. Nazar are with the School of Mathematical Science,
Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor, Malaysia (email: anuarishak@yahoo.com, rmn72my@yahoo.com).

Issue 3, Volume 4, 2010

been considered by Char and Chen [9], Lin and Cheng [10],
Elbashbeshy [11] and very recently by Ishak et al. [12].
All of the above mentioned studies dealt with stretching
sheet where the flows were assumed to be steady. The
unsteady flows due to a stretching sheet have received less
attention; a few of them are those considered by Devi et al.
[13], Andersson et al. [14], Nazar et al. [15], and very recently
by Ishak et al. [16]. In Ref. [15], the similarity transformation
introduced by Williams and Rhyne [17] was used, which
transformed the governing partial differential equations with
three independent variables to two independent variables,
which are more convenient for numerical computations.
Motivated by the above investigations, in this paper we
present the unsteady flow and heat transfer characteristics
caused by a stretching sheet immersed in a micropolar fluid.
The governing partial differential equations are transformed
into ordinary ones using similarity transformation, before
being solved numerically by the Keller-box method. The
results obtained are then compared with those of Elbashbeshy
[11] and the series solution for the steady-state flow case to
support their validity.
II. PROBLEM FORMULATION
Consider an unsteady, two-dimensional laminar flow of an
incompressible micropolar fluid over a stretching sheet. At
time t = 0 , the sheet is impulsively stretched with velocity
U w ( x, t ) along the x -axis, keeping the origin fixed in the fluid
of ambient temperature T∞ . The stationary Cartesian
coordinate system has its origin located at the leading edge of
the sheet with the positive x -axis extending along the sheet,
while the y -axis is measured normal to the surface of the
sheet. The boundary layer equations may be written as [12,16]
∂u ∂v
+
=0,
∂x ∂y

∂u
∂u
∂u  µ + κ  ∂ 2 u κ ∂N
,
+u
+v
=
+

∂t
∂x
∂y  ρ  ∂y 2 ρ ∂y
 ∂N

∂N
∂N 
∂2 N
∂u 
ρ j
+u
+v
− κ  2N +  ,
=γ
2

t

x

y
∂y 
∂y


∂T
∂T
∂T
∂2T
+u
+v
=α 2
∂t
∂x
∂y
∂y

167

(1)
(2)
(3)
(4)

Thus. spin gradient viscosity. Eqs. the spin-gradient viscosity γ can be defined as 1/ 2 y. f (0) = 0. T → T∞ as y → ∞. b. b ≥ 0 and c ≥ 0 (with ct < 1 ). b and c are constants with a > 0 . z ) denotes the confluent hypergeometric function [21]. t ) have been chosen in order to be able to devise a new similarity transformation. − Pr e e Pr M ( Pr − 1. =− w ∂y ∂y k u → 0. vortex viscosity. q w ( x. µ. It is assumed that the stretching velocity U w ( x. respectively. (12) The quantities of physical interest are the values of f ′′(0) and f (η ) = 1 − e − η . the skin friction coefficient f ′′(0) and the surface temperature θ (0) can be shown to be given by f ′′(0) = −1 . t ) and the surface heat flux qw ( x. It should be noted that at t = 0 (initial motion). thereby facilitating the exploration of the effects of the controlling parameters (see Andersson et al. and α are the microinertia per unit mass. (5) where m is the boundary parameter with 0 ≤ m ≤ 1 [18]. h(0) = −mf ′′(0). Pr.and y directions. where η is the similarity variable. 12  f ′(0) = 1. (13) and (14). These particular forms of U w ( x. ψ = (ν xU w ) 1/ 2 1/ 2 U  N = Uw  w  ν x  f (η ) . Relation (6) is invoked to allow the field of equations predicts the correct behavior in the limiting case when the microstructure effects become negligible and the total spin N reduces to the angular velocity [19.26]). The boundary conditions (5) now become (7) where K = κ/µ is the dimensionless viscosity ratio and is called the material parameter. T is the fluid temperature in the boundary layer. h (η ) → 0. ∞ The continuity equation (1) is satisfied by introducing a stream function ψ such that u = ∂ψ / ∂y and v = −∂ψ / ∂x . The transformed nonlinear ordinary differential equations are: 1 (1 + K ) f ′′′ + ff ′′ − f ′2 + Kh ′ − S  f ′ + η f ′′  = 0 Issue 3. angular momentum and energy equations can be transformed into the corresponding ordinary differential equations by the following transformation:  Uw   ν x  where primes denote differentiation with respect to η. t ) and qw ( x. z ) = 1 + ∑ an = a ( a + 1)(a + 2) ⋯ (a + n − 1) . As was shown by Ahmadi [19]. [23. − Pr ) . θ (0) = (15) 1 M ( Pr − 1. our task is to investigate how the governing parameters S . and both a and c have dimension time −1 . t ) = 1 − ct 1 − ct where a . Pr = ν/α is the Prandtl number and S = c / a is the unsteadiness parameter. − Pr ) (16) The nonlinear ordinary differential equations (9) – (12) have been solved numerically by a finite-difference scheme known as the Keller-box method. m. θ ′(0) = −1. ρ. Pr + 1. N → 0. N = −m . We note that when K = 0 (viscous fluid) and S = 0 (steady flow). K and Pr influence these quantities. θ (η ) = (13) 1 −η Pr M ( Pr − 1. Pr. at y = 0. Pr M ( Pr − 1.INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES 1  K  ′′ 3  1 +  h + f h ′ − f ′ h − K ( 2h + f ′′ ) − S  h + η h ′  = 0 2 2 2     1 1   θ ′′ + f θ ′ − f ′θ − S  θ + ηθ ′  = 0 Pr 2   subject to the boundary conditions q ∂u ∂T u = U w . N is the microrotation or angular velocity. θ (η ) → 0 as η → ∞. with 2  an z n . t ) = . the problem under consideration reduces to a steadystate flow. γ . bn = b(b + 1)(b + 2) ⋯ (b + n − 1) . fluid density and thermal diffusivity. η = (11) 1 θ (0) which represent the skin friction coefficient and the heat transfer rate at the surface. which transforms the governing partial differential equations (1) – (4) into a set of ordinary differential equations. which is very familiar to the present authors (see Bachok et al. as described in the book by Cebeci and Bradshaw [22]. By using Eqs. t ) are of the forms ax bx U w ( x. b. Pr + 1. κ. u and v are the velocity components in the x .24] and Bachok and Ishak [25. 2010 f ′(η ) → 0. respectively.20]. respectively. (8) k ( T − T∞ )  U w  h (η ) . (10) III. θ (η ) =   qw  vx  . Volume 4. v = 0. n =1 bn n ! M ( a . SOLUTION PROCEDURE (9) (i) Finite-difference method 168 . [14]). The momentum. and j . dynamic viscosity. (14) where M ( a. − Pr ) −η ). where the closed-form solution for the flow field and the solution for the thermal field in terms of Kummer’s functions are respectively given by (6) γ = ( µ + κ / 2 ) j = µ (1 + K / 2 ) j . (1) – (4) describe the steady flow over a stretching surface.

Rearranging of expressions (26)-(32) gives (17) (18) 1 f j − f j −1 − h j ( p j + p j −1 ) = 0. Volume 4. p (jk +1) = p (jk ) + δ p(jk ) . s (η ) → 0 as η → ∞. p0 = 1. 2010 )( )( ( + K ( n ) j −1 2 − 1   S ( p ) j −1 2 + η ( q ) j −1 2  = 0 2    K  n j − n j −1 + ( f n ) j −1 2 − ( p g ) j −1 2 − 1 +  hj 2  (36) S number that indicates the coordinate location. g (η ) → 0.spacing and j = 1. and the difference equations are then expressed using central differences. q (j (31) k +1) = q (j ) + δ q (j ) .. 2  2  K  2 ( g ) j −1 2  ( ( (37) (38) ) ) ) ( (39) ) Equations (33)-(39) are imposed for j = 1. (33) (1 + K ) q j − q j −1 + (23) In terms of the new dependent variables. η J = η∞ . (25) hj p j − p j −1 hj g j − g j −1 hj s j − s j −1 hj (1 + K ) p j + p j −1 2 = = = q j + q j −1 2 n j + n j −1 2 t j + t j −1 2 q j − q j −1 hj = p j −1 2 . s (jk +1) = s (jk ) + δ s (jk ) . s J = 0. g 0 = −mq0 . 2 (19) (20) 1 (1 + K ) q ′ + fq − p 2 + Kn − S  p + η q  = 0 . p (η ) → 0. The boundary conditions are f 0 = 0. (9) – (11) are first converted into a system of seven first-order equations. Pr 2   2 (21) (22) p (0) = 1. 2 1 s j − s j −1 − h j ( t j + t j −1 ) = 0. (26) = t j −1 2 .. (27) = n j −1 2 . s ′ = t. by introducing the following expression: f j( k +1) = f j( k ) + δ f j( k ) . g (0) = − mq (0). 2 1 g j − g j −1 − h j ( n j + n j −1 ) = 0.   1   K 3 1 +  n ′ + f n − p g − K ( 2 g + q ) − S  g + η n  = 0 . and the (29) transformed boundary layer thickness η J is to be sufficiently large so that it is beyond the edge of the boundary layer. p ′ = q. The finitedifference approximation equations (17)-(23) are written for the midpoint η j −1 2 of the segment η j −1 η j . the boundary conditions (12) are given by f (0) = 0. we use Newton’s method. For this purpose. + ( fq ) j −1 2 − ( p 2 ) j −1 2 1 3 1  h j  ( g j + g j −1 ) + η ( n j + n j −1 )  = 0 2 2 2  1 1 t j − t j −1 + h j f j + f j −1 t j + t j −1 − Pr 4 1 h j p j + p j −1 s j + s j −1 − 4 1  1  S h j  s j + s j −1 + η t j + t j −1  = 0. q(η ) . 2   t (j 169 k +1) = t (j ) + δ t (j k k) (41) . which is defined as below: η0 = 0. k k n(jk +1) = n(jk ) + δ n (jk ) . η j = η j −1 + h j . 2. (26) = q j −1 2 . g (j k k k +1) = g (j ) + δ g (j ) . t (0) = −1. t0 = −1.. 1 3  + ( q ) j −1 2  − S  ( g ) j −1 2 + η ( n ) j −1 2  = 0  2 2   Issue 3. s(η ) = θ (η ) and t (η ) so that Eqs. (9) – (11) can be written as f ′ = p. J . J is a sequence gives f j − f j −1 (34) 1 h j ( f j + f j −1 )( q j + q j −1 ) − 4 2 1 1 h j ( p j + p j −1 ) + h j K ( n j + n j −1 ) − 4 2 1 1   h j S ( p j + p j −1 ) + η ( q j + q j −1 )  = 0 2 2   We now consider the segment η j −1 η j .. . 2 2 2     1 1   t′+ f t − p s − S  s + η t  = 0 . 3. g (η ) = h (η ) . (30) (ii) Newton’s method To linearize the nonlinear system (33)-(39).INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES 1 t j − t j −1 t ′ + ( f t ) j −1 2 − ( p s ) j −1 2 − Pr h j To solve the transformed differential Eqs.. with η j −1 2 as the midpoint. 2.. Eqs. g J = 0. we introduce new dependent variables p(η ) . g ′ = n. This procedure = (35) 1  K 1 +  n j − n j −1 + h j ( f j + f j −1 )( n j + n j −1 ) − 2 4  1 h j ( p j + p j −1 )( g j + g j −1 ) − 4 1 K h j  2 ( g j + g j −1 ) + ( q j + q j −1 )  − 2 (24) where h j is the ∆η . 2 1 p j − p j −1 − h j ( q j + q j −1 ) = 0.. (32) 1   S ( s ) j −1 2 + η ( t ) j −1 2  = 0. (9) – (11) subjected to the boundary conditions (12). (40) pJ = 0. n(η ) .

δ t0 = 0. ( a8 ) j = ( a7 ) j . (50) η   Sh j ( p ) j −1 2 + ( q ) j −1 2  . In a matrix-vector form. 2 (51) which just express the requirement for the boundary conditionts to remain constant during the iteration process. 2 K 1 S ( b1 ) j = 1 + + h j f j −1 2 − h jη . δ n( k ) . (iii) Block-elimination method The linearized difference equations (42)-(48) can be solved by the block-elimination method as outlined by Cebeci and Bradshaw [22]. δ p0 = 0. ( b9 ) j = − h j . ( c3 ) j = h j t j −1 2 . 2 1 S = ( c5 ) j . ( a3 ) j ( a4 ) j = ( a3 ) j . (42)-(48) can be written as 1 = − h j g j −1 2 . ( r3 ) j −1 2 = − g j + g j −1 + h j n j −1 2 . (43) j + δ n j −1 ) = ( r3 ) j −1 2 . δ p ( k ) . = A =r where 170 (52) . 2 2 4 K 1 ( b2 ) j = ( b1 ) j − 2 1 +  . 2 4 K = ( b7 ) j . δ q ( k ) . Pr 2 1 = ( c3 ) j . 2 ( b4 ) j = ( b3 ) j . ( b5 ) j ( b6 ) j = ( b5 ) j . (42) j + δ q j −1 ) = ( r2 ) j −1 2 . but here an interesting feature can be observed that it consists of block matrices.1. ( c2 ) j δ s ( k ) and δ t ( k ) . 2 4 ( a2 ) j = ( a1 ) j − 2 (1 + K ) . 1 3 = − h j p j −1 2 − Kh j − Sh j . We then insert the left-hand side expressions in place of δ f ( k ) . Pr 2 4 2 1 = ( c1 ) j − . 2 Issue 3. . δ g0 = 0. δ pJ = 0. ( b1 ) δ n j + ( b2 ) δ n j −1 + ( b3 ) δ f j + ( b4 ) δ f j −1 + ( b5 ) δ p j + ( b6 ) δ p j −1 + ( b7 ) δ g j + ( b8 ) δ g j −1 + ( b9 ) δ q j + ( b10 ) δ q j −1 = ( r6 ) j −1 2 . 2010 − Kh j ( n ) j −1 2 + (49) δδδδ ( b8 ) j j −1 2 = 1 ( a6 ) j = ( a5 ) j . 2. 2 2 = ( c7 ) j .  2  1 = ( t j − t j −1 ) − h j ( ft ) j −1 2 + Pr η   h j ( ps ) j −1 2 + Sh j ( s ) j −1 2 + ( t ) j −1 2  . 2 2  ( b7 ) j 1 1 S + h j f j −1 2 − h jη . ( c5 ) j = − h j s j −1 2 . ( b10 ) j = ( b9 ) j .. δ sJ = 0. δ g J = 0. ( a5 ) j = −h j p j −1 2 1 h j q j −1 2 . This procedure yields the following linear ( c4 ) j system (the superscript k is dropped for simplicity): hj (δ p 2 δ f j − δ f j −1 − + δ p j −1 ) = ( r1 ) j −1 2 . (45) (46) ( r5 ) j −1 2 = (1 + K ) ( q j − q j −1 ) − h j ( fq ) j −1 2 + hj ( p2 ) (47) ( r6 ) j −1 2 (48) where ( a1 ) j ( r7 ) j −1 2 = 1+ K + 1 S h j f j −1 2 − h jη . (44) hj (δ q 2 δ p j − δ p j −1 − hj (δ n 2 δ g j − δ g j −1 − δ s j − δ s j −1 − j hj (δ t 2 j + δ t j −1 ) = ( r4 ) j −1 2 . 2   The boundary conditions (40) become δ f 0 = 0.INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES ( c1 ) j where k = 0. δ g ( k ) . ( a1 ) δ q j + ( a2 ) δ q j −1 + ( a3 ) δ f j + ( a4 ) δ f j −1 + ( a5 ) δ p j + ( a6 ) δ p j −1 + ( a7 ) δ n j + ( a8 ) δ n j −1 = ( r5 ) j −1 2 . Eqs. Volume 4. ( c6 ) j ( c8 ) j and ( r1 ) j −1 2 = − f j + f j −1 + h j p j −1 2 . the block-tridiagonal structure consists of variables or constants. ( c1 ) δ t j + ( c2 ) δ t j −1 + ( c3 ) δ f j + ( c4 ) δ f j −1 + ( c5 ) δ p j + ( c6 ) δ p j −1 + ( c7 ) δ s j + ( c8 ) δ s j −1 = ( r7 ) j −1 2 . ( r4 ) j −1 2 = − s j + s j −1 + h j t j −1 2 . ( c7 ) j = − h j p j −1 2 − h j . Commonly. 2    K = 1 +  ( n j − n j −1 ) − h j ( fn ) j −1 2 + 2  h j ( gp ) j −1 2 + Kh j  2 ( g ) j −1 2 + ( q ) j −1 2  +   S  h j 3 ( g ) j −1 2 + η ( n ) j −1 2  . ( a7 ) j = h j K . 2 S − hj . ( r2 ) j −1 2 = − p j + p j −1 + h j q j −1 2 . since the system has block-tridiagonal structure. .. ( b3 ) j = h j n j −1 2 .

 δ q1     δ n1  δt   1 2 ≤ j ≤ J. we assume that A is nonsingular and it can be factorized as A = LU .  r   ( 5 ) j −1 2  ( r )   6 j −1 2  ( r7 )  j −1 2   (58) To solve Eq. δ q0  δ n   0  δ s0    [δ1 ] = δ f1  .  ( r1 ) j −1 2    ( r2 ) j −1 2     ( r3 ) j −1 2     rj  = ( r4 ) j −1 2  .  1  − 2 hJ   −1    0 [ AJ ] =   0   (a )  6 J  ( b6 ) J   ( c6 ) J −1 1 ≤ j ≤ J − 1.     [δ J −1 ] [ rJ −1 ]  [δ ]   [r ]   J   J  The elements of the matrices are as follows: 0 0  0  1  − h1 0 0  2  1  0 − h1 0 2  [ A1 ] =  1 0 − h1  0 2   ( a2 )1 ( a8 )1 0  0  ( b10 )1 ( b2 )1  0 0 c ( 2 )1  1 0 0 0 0 1 − h1 2 0 0 0 0 ( a3 )1 ( a1 )1 ( a7 )1 ( b3 )1 ( b9 )1 ( b1 )1 0 ( c3 ) 1 0 0 0 1 0 0 0 0 0 1 − hJ 2 0 −1 0 0 0 1 − hJ 2 0 −1 0 0 0 0 ( b8 ) J 0 ( a10 ) J ( a3 ) J ( a1 ) J 0 ( b3 ) J ( b9 ) J c 0 ( 8 ) J ( c3 ) J 0 0 0 0 0 0 0 1 − hJ 2 0 0 0 0 0 0 0 0 0 0 0 0   0    0   1  − h1  2  0   0  ( c1 ) 1   1  − 2 hJ   1  0 [CJ ] =  0  a  ( 5 )J  ( b5 ) J   ( c5 ) J (53) 0 1 − hJ 2 ( a4 ) J ( a2 ) J ( b4 ) J ( b10 ) J 0 ( c4 ) J 0 0 ( a8 ) J ( b2 ) J 0   0    0   1  − hJ  2  0   0  ( c2 ) J  0 (55) ( a7 ) J ( b1 ) J 0 0 0 0 0 1 0 0 0 1 ( a9 ) J ( b7 ) J 0 0 ( c7 ) J  0 0 0 0  0 0 0 0 0 0 0 0  0 0 0 0 0 0 0 0  0 0 0 0  0 0 0 0  (56) 2 ≤ j ≤ J. Issue 3. 2010 171 (59) . (52). 1 − h1 2 0 0 0  0   0    0   (54) 1  − hJ 2  0   0  ( c1 ) J  δ q j −1  δ n   j −1  δ s j −1    δ j  =  δ f1   δ qj     δ nj   δt   j  (57) and 1 ≤ j ≤ J .INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES  [ A1 ]  [ B2 ]   A=      [C1 ] [ A2 ] [C2 ] 0  0   0 [ BJ ] =  0  0  0  0         [CJ −1 ] [ AJ ]  ⋱ ⋱ ⋱ [ BJ −1 ] [ AJ −1 ] [ BJ ]  [δ1 ]   [ r1 ]       [δ 2 ]   [ r2 ]  δ =  ⋮  and r =  ⋮  . Volume 4.

[δ i ] = W j  − Γ j  δ j +1  .01 was found to be satisfactory for a convergence Eq. This is probably because in boundary layer calculations. j = 2. in nearly all cases. unconditionally stable and is easy to be programmed. Eqs. (64) W. a uniform grid of ∆η = 0. These      [Γ J −1 ] [ I1 ]  calculations are repeated until the convergence criterion is satisfied. and W j  are 7 × 1 column matrices. the wall shear stress parameter q ( 0 ) is commonly used as the convergence criterion (Cebeci and Bradshaw [22]). (61) (62) (63) Substituting Eq. The results are given to carry out a parametric study showing the influences of the unsteadiness parameter S . (65) LU (71) (60) If we define = δ U IV. [28. as well as the series solution given by Eq. where ∈1 is a small prescribed value. in order that the numerical values obtained are mesh independent. 3. However. (52). we obtain = δ r. [α i ] Γ j  = C j  . (65) by the following relations: [α1 ][W1 ] = [ r1 ] . . j = 2. . LW (66) where ====  [W1 ]     [W2 ]   ⋮ . the boundary-layer thickness η∞ between 5 and 50 was chosen where the infinity boundary condition is achieved. Volume 4. δ q0( k ) <∈1 .. J . ∈1 = 0. The elements of W can be determined from Eq. 3. at least to four decimal places. this convergence criterion is used in the present study. it is found that the greatest error usually appears in the wall shear stress parameter. In this study. (68) When the elements of W have been found. On the other hand. the values of ∆η between 0. (41).. where [α1 ]  [ β 2 ] L=     [α 2 ] ⋱ ⋱ [α J −1 ] [βJ ]       [α J ] 2 ≤ j ≤ J.. thus making it highly attractive for production use.1 were used. J − 1.. (70) and [ I1 ]   U =     [Γ1 ] [ I1 ] ⋱ ⋱ [ I1 ] Once the elements of δ are found. RESULTS AND DISCUSSION The step size ∆η in η .INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES [α i ] W j  =  rj  −  B j  W j −1  .29]. while [α i ] and [ Γi ] are 7 × 7 matrices in which elements are determined by the following equations: [α i ] = [ A1 ] . (42)-(48) can be used to find the ( k + 1) th interation in Eq. which gives about four decimal places accuracy for most of the predicted quantities as suggested by Bachok and Ishak [25. depending on the values of the parameters considered. (64) becomes ==== r.00001 is used.27] and Ali et al. (65) gives the solution for δ in which the elements are found from the following relations: [δ J ] = [WJ ] . In this study.. Thus. [ A1 ][Γ1 ] = [C1 ] . The present method has a second-order accuracy. (59) into Eq. [αi ] = [ A1 ] −  B j  [Γ J −1 ] . (69) 1 ≤ j ≤ J − 1.001 and 0.. In laminar boundary layer calculation.   [WJ -1 ]  [W ]   J  W criterion of 10−5 which gives accuracy to four decimal places. Issue 3. material parameter K and Prandtl number Pr . Calculations are stopped when where [ I ] is a 7 × 7 identity matrix. 2010 (67) 172 . The only disadvantage is the large amount of once-and-for-all algebra needed to write the difference equations and to set up their solutions. Eq. the case when S = 0 (steady-state flow) has also been considered and the results are compared with those of Elbashbeshy [11]. and the position of the edge of the boundary-layer η∞ have to be adjusted for difference values of the parameters to maintain the necessary accuracy. For validation of the numerical method used in this study.

(16) Numerical results 0 0. the heat transfer rate at the surface 1 θ (0) increases with Pr . Figures 3-6 show the temperature profiles for selected values of parameters.9 3 4 5 6 0 0 7 η FIG.2 0. Figure 7 shows that the surface temperature θ (0) decreases with increasing values of K. 2 0. As can been seen from Figs.2 0.5. K and Pr .INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES (16). 1 Velocity profiles f ′ (η ) for some values of S when 2 4 6 η 8 10 12 FIG. 5). the microrotation at the surface h ( 0 ) is more dominant for larger values of m .5. Figs.8591 1 1 Pr = 1.72 1 1. which represent the heat transfer rate at the surface TABLE 1 Variations of θ (0 ) for different values of S and Pr S Pr Elbashbeshy [11] Eq. 3.9 f ′ (η) f ′ (η) 0. This is because the higher Prandtl number fluid has a lower thermal conductivity (or a higher viscosity) which results in thinner thermal boundary layer and hence.5 S = 0. the local Nusselt number 1 θ (0) . 0. when the other parameters are fixed to unity is presented in Fig. the surface temperature θ (0) decreases when Pr is increased.7 0. The temperature profiles are found to subside monotonously to zero as η increases.72 1 0. 2 Velocity profiles f ′ (η ) for some values of K when Pr = 1. Thus. increasing S .236657472 1. micropolar fluids show drag reduction compared to viscous fluids. Figure 2 shows that the velocity gradient at the surface is larger for larger values of S .9116 0. Thus. The quantitative comparison is shown in Table 1 and it is found to be in a very good agreement. 0. S = 1 0.7 0. m = 0. The velocity gradient at the surface f ′′ ( 0 ) decreases as K increases. 1.5. 2010 K = 0. It is evident from Figure 2 that the boundary layer thickness increases with K . m = 0. 3-6. i. Thus.4 0. The effect of m on the angular velocity. m = 0.6 0. 8.5 and S = 1 173 . (9)-(11). thus supporting the numerical results obtained.3 0.e. On the other hand. 2.6 0. The velocity profiles for various values of S and K are presented in Figs.5 and K = 1 Issue 3. K = 1 0. higher heat transfer rate at the surface (see Fig. which is clear from Eqs. We note that the parameter Pr have no influence on the flow field.5.1 0 0 1 2 Pr = 1.8 0.4 0.5 0. m = 0. for a fixed value of K . 1.2253 1. These curves represent the physically realistic case.0000 1 0. K or Pr increases.2367 1.3 0.1 0.000000000 1. 1-6 show that the far field boundary conditions (12) are satisfied asymptotically. the surface temperature θ (0 ) decreases with increases when S . 1 and 2. As expected. Volume 4.0000 1. the heat transfer rate at the surface 1 θ (0) is higher for a micropolar fluid ( K > 0 ) compared to a Newtonian fluid (K = 0). which produces larger skin friction coefficient f ′′ ( 0 ) . 4 Pr = 1. Finally.8 0.

1 0 0 1 2 3 4 5 η 6 7 8 9 0 0 10 2 4 6 8 10 η FIG. 5 0. 6 Temperature profiles θ (η ) for some values of K FIG.1.8 0.4 Pr = 0. m = 0.5 0. 4 Temperature profiles θ (η ) for some values of S when Pr = 0.5 and K = 1 Issue 3. m = 0.7.5.6 0.2 0. K = 2 Pr = 1. 0.5 and K = 2 when Pr = 1.5.8 0. 5 Temperature profiles θ (η ) for some values of Pr when Pr = 1.8.7. 0.7 0.3 K = 0.7 0. m = 0.5.2 Pr = 1. K = 1 1 0.2 0.1 when S = 0.5.7 .6 0. 0.5 and S = 0. 3 Temperature profiles θ (η ) for some values of S FIG. m = 0. 1.9.2 1 Pr = 0. 0.1 0 0 1 2 3 η 4 5 0 0 6 4 6 8 η FIG.9 S = 0.9 0. Volume 4.2 0. m = 0.2 0.5. 0.4 0.INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES 1. m = 0.5 0.5.6 0.9 1 0. 2010 2 174 10 .8 θ (η ) θ(η) 0.8 θ(η) θ (η ) 0.7.4 0. 1 0.4 S = 0. 0. m = 0. 0. K = 1 0.3 S = 0.5 and K = 1 1 1.1.9 0. m = 0. S = 0.6 0.1 0.

1 m when S = 1.7 0. vol. which transform the time-dependent governing equations to ordinary differential equations. 2010 [13] C.5. N. I. Char and C. 6 – 16. M. Crane. 69-74. Amin and I. Heat Mass Transfer. Chen.1 1 2 3 4 0 0 5 K 1 2 3 4 5 η FIG.0 0. Ishak. “Unsteady mixed convection flow in stagnation region adjacent to a vertical surface. We discussed the effects of the governing parameters S. Keller. 2004. [14] H.9 Pr = 1. Dandapat. 175 . pp.” SIAM J. Chen. J. S. 1998. [16] A. 577-585. 471-476. Pop. 1988.” Int. Magyari and B. Appl. 1970. 2. 121-128. Eng. S. 31. “Flow past a stretching plate.C.” Mech. J.6 1 h(η ) θ(0) 1.2 0. [8] E. Ishak. 33. R. vol. vol. Lin and C.9 S = 0. 215-224. 33.J. Y. 559-561. Phys. 55. Nazar. Nazar and I. continuously stretching sheets.” Heat Mass Transfer. 31. Takhar and G. [6] C. H. B. D: Appl.2 0. “Approximate analytic solutions for mixed and forced convection heat transfer from an unsteady no-uniform flow past a rotating cylinder. 2009.” J.” J. Pr = 1. 1999. Bobba.” Heat Mass Transfer. [10] C.INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES 1 1. vol. Andersson. K and Pr.5 1. K. 1980. pp. Commun. [3] L. “ Unsteady boundary layer flow due to stretching surface in a rotating fluid. Res. the heat transfer rate at the surface is higher for a micropolar fluid compared to a Newtonian fluid. “ Can. 1998.. pp. “ WSEAS Transactions on Heat and Mass Transfer.3 0. 645 – 647.1 0. 372. “ Unsteady mixed convection heat and mass transfer past an infinite porous plate which thermophoresis effect. J. [1] K. I. vol. 10 0. S. 38.4 0. vol. pp. 31. 107. as well as the series solution for the steady-state flow. D: Appl. I. Further. Phys. R. Abdell and F. 4. 477-480. pp. Rhyne. [15] R. 5. Math. We found that the heat transfer rate at the surface 1 θ (0) increases with S.4 1. Phys. [11] E. vol.7 0. Phys. “Exact solution of heat transfer from stretching surface with variable heat flux. CONCLUSIONS [7] C. Pr = 1 and K = 1 V. “Laminar mixed convection adjacent to vertical. Aarseth and B. Chen.7 Pr = 0. REFERENCES [12] A.. vol. 26. 43. M. “ Temperature field in non-Newtonian flow over a stretching plate with variable heat flux. Volume 4. 1998. H. 71-79. pp. “Heat transfer characteristics of a continuous stretching surface with variable temperature. “ Boundary layer development on a wedge impulsively set into motion. [5] L. [9] M. vol. Pop. A.” Phys. (ZAMP). K and Pr on the fluid flow and heat transfer characteristics. Al-Azab. [2] T.8 m = 0. 1977. Devi. vol. vol.7 and FIG. vol. Nath. “ The three-dimensional flow due to a stretching surface. Heat Mass Transfer. vol. 1984.1 0. pp.5 0. [17] J. Appl.8 1.B. 7 Variations of θ ( 0 ) with K for Pr = 0. pp. The numerical results compared very well with previously reported cases. 744746. “ Heat and mass transfer on a stretching sheet with suction or blowing. 10. pp. K. 23-33. vol. pp. 19511954. [4] P.0 when S = 0. pp. 32. 1991. Chem. S. S. A. Nazar an.” Phys. 2008. Pop. vol. Gupta. 21. Williams III and T. “ Heat transfer over an unsteady stretching permeable surface with prescribed wall temperature. pp. RWA.” Nonlinear Anal.” ASME J. 2009. pp. K = 1 0. “Heat transfer in a liquid film on an unsteady stretching surface. Fluids. We have theoretically studied the similarity solutions of the unsteady boundary layer flow and heat transfer due to a stretching surface.” Int. 248-250. Vol. Magpantay. pp.” Heat and Mass Transfer. 0.5 0 S = 1. D. 2909-2913.” WSEAS Transactions on Heat and Mass Transfer. 1985. A new similarity solution has been devised. 8 Angular velocity profiles h (η ) for some values of Pr = 1. vol. pp. pp. 27. Heat Transfer. “ Heat transfer over a stretching surface with variable heat flux in micropolar fluids.” J. R. Elbashbeshy. Issue 3.3 0. J. Math. 2000. Grubka and K. Gupta and A. J. pp. 917-921. Phys. 2. “Heat transfer over a stretching surface with variable surface heat flux. 2007. Wang.6 0. “ Heat and mass transfer in the boundary layers on an exponentially stretching continuous surface. 1915-1917. 1. Lett.

Bachok. M. 1965. 2010. pp. J. M. pp. Ishak and I. 2010 176 . on Appl of Comp Eng. 411-418. pp. 2010. Nazar and N. 5. “ Physical and Computational Aspects of Convective Heat Transfer”.” MHD viscous flow and heat transfer induced by a permeable shrinking sheet with prescribed surface heat flux. [20] A. A. R. vol.” Flow and heat transfer over a stretching cylinder with prescribed surface heat flux. J. 82. 49. 14. 1225-1236.” Boundary-layer flow of nanofluids over a moving surface in a flowing fluid. New York. 1988. Nazar and I. pp. 639-646. [26] N. 2010. [21] M. 43. 260-263.” Proceedings of the 9th WSEAS Int. Ali. on Appl of Comp Eng.” Proceedings of the 9th WSEAS Int. vol. Pop. Bachok and A. Ishak and I. M. vol. [18] A.” Int. 363-373. J. of Thermal Sciences. Nazar and I. Abramowitz and I.. Ishak. New York.A. 115118. pp. Bradshaw.” Mixed convection boundary layer flow near the stagnation point on a vertical surface embedded in a porous medium with anisotropy effect. [22] T. Dover.” Flow and heat transfer characteristic on a moving flat plate in a parallel stream with prescribed surface heat flux. Conf. 73-82. R. 4. 365-375. 1663-1668. Sci. Ishak. [28] F.INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES [25] N. vol. [29] F.” WSEAS Transactions on Mathematics. 1976. 2010.” Mixed convection stagnation point flow of a micropolar fluid towards a stretching sheet. Ishak. pp. Eng. 2010. Arifin. R. pp. Ishak. Sci. pp. [24] N. Volume 4.” Int. Cebeci and P. 44. 159-169. pp. Issue 3. Bachok and A. A. [19] G. “Self-similar solution of incompressible micropolar boundary layer flow over a semi-infinite plate. Conf. Nazar and N. vol.” Trans. “Handbook of Mathematical Functions”. [27] N. Ali.. [23] N. vol. Eng. Ishak. Arifin. pp. Porous Med.” Malaysian Journal of Mathematical Sciences. Bachok and A. 2006.” MHD viscous flow and heat transfer due to a permeable shrinking sheet with prescribed surface heat flux. R.” The effects of suction and injection on a moving flat plate in a parallel stream with prescribed surface heat flux. Springer. Stegun.” Int. vol. Pop. Pop. 2008. Ahmadi. Pop. Bachok. 2010. M. 2010.” Meccanica. “Moving wedge and flat plate in a micropolar fluid.” WSEAS Transactions on Heat and Mass Transfer.