INTRODUCTION
The study of laminar flow in a boundary layer caused by a moving rigid surface was initiated by Sakiadis (1961). Later the work was extended by Crane (1970) to the flow due to stretching of a sheet. Heat transfer in the boundary layer of an incompressible fluid past a continuous moving surface has several industrial applications in textile and paper industries, extrusion process of plastic sheets, spinning of fibers etc. NonNewtonian fluids have since gained considerable importance due to their extensive use in industry. Viscoelastic boundary layer flow over a stretching sheet has been the main subject of a number of researchers in the past (Rajagopal and Gupta, 1984, Rollins and Vajravelu, 1992). It has been generally found that heat transfer in the visco elastic fluid is less compared to that of a Newtonian fluid. Viscoelastic fluids are therefore more popularly used in industry than Newtonian fluids.
The transport of heat in a porous medium has also considerable practical applications in geothermal systems, crude oil extraction, ground water pollution and also in a wide range of biomechanical problems.
In the earlier investigations involving heat transfer, two different heating processes have been considered namely 1. Prescribed Surface Temperature (PST case) and 2. Prescribed Heat Flux (PHF case). Vajravelu (1994) studied the flow of a steady viscous fluid and heat transfer characteristics in a porous medium by considering both these cases. Abel et al. (1998) studied the viscoelastic fluid flow and heat transfer characteristics in a saturated porous medium over an impermeable stretching surface. Abel et al. (2004) carried out the study of the momentum, mass and heat transfer characteristics on the flow of a viscoelastic fluid past a stretching sheet in the presence of a transverse magnetic field.
The influence of radiation on hydrodynamic flow and heat transfer characteristics is also recognized by some investigators. Raptis (1999) studied the effect of radiation on viscoelastic flows. The magneto hydrodynamic flow and heat transfer of a viscoelastic fluid over a stretching sheet considering the radiation effects has been studied by Siddheshwar and Mahabaleswar (2005). The study of the effects of variable viscosity and variable thermal conductivity on heat transfer from a moving surface in a micropolar fluid through a porous medium considering radiation is done by Elsayed Elbarbary et al. (2004).
In all the above cases involving radiation, the studies were essentially based
on closed form analytical solutions. The aim of the present study is to use
a numerical technique to study the effect of radiation on the viscoelastic
flow and heat transfer in a porous medium over a stretching sheet and compare
the results with those of earlier researchers. For the heat transfer category
the two cases namely PST and PHF have been considered. The governing partial
differential equations have been reduced to a system of ordinary differential
equations by similarity transformation. This equation set is then solved numerically
by the RungeKutta method using the shooting technique for the relevant boundary
conditions and flow parameters. The results obtained for typical values of the
influencing parameters in nondimensional form which include the viscoelastic
parameter (K_{1}), the permeability parameter (K_{2}), radiation
number (N_{R}), the Prandtl number (Pr) and heat source/sink parameter
(β) are graphically represented to show the degree of influence of these
variables on the flow and the heat transfer characteristics. These parameters
will be defined in subsequent sections.
MATHEMATICAL FORMULATION
Consider a steady twodimensional flow of an incompressible viscoelastic fluid (Walter’s liquid B model) through a porous medium past a semiinfinite stretching sheet. The following notations are used in the mathematical treatment that follows:
Let x and y represent the tangential (along flow) and transverse (across flow) directions respectively; let u and v denote the velocity components of the fluid in the x and y directions. T the temperature, v the kinematic viscosity, k_{0} is the coefficient of viscoelasticity, k’ is the permeability coefficient of the porous medium and ρ is the fluid density. Further, let k denote the thermal conductivity, c_{P}, the specific heat at constant pressure and q_{r}, the radiative heat flux. Let Q represent the volumetric rate of heat generation, T_{w} the sheet surface temperature, T_{∞} the temperature at the farthest location representing the edge of the boundary and q_{w} denote the heat transfer rate per unit area. λ is the stretching rate of the sheet (assumed constant) along x axis.
The fluid flow is confined to the half space y>0. By applying two equal
and opposite forces along the xaxis, the sheet gets stretched with a speed
proportional to the distance from the fixed origin i.e., u = λx. The continuity
equation and the basic boundary layer equations governing the flow and heat
transfer including radiation take the form of Eq. 1 to 3.
Continuity equation:
Equation governing velocity of flow:
Equation governing heat transfer:
Boundary conditions: For the PST case, the surface temperature and for the PHF case, the heat flux are considered as a power series in x. The expressions for temperature T involve the characteristic length l and two distinct constants A and D for the PST and PHF cases, respectively.
(i) 
At y  0:u = λx, v = 0 (λ>0) 
(ii) 
As y→∞: u→0, u_{y}→0 
Transformation of equations: The continuity Eq. 1 is satisfied by the stream function defined by
The similarity transformation is introduced in the form
The nondimensional temperature θ and g for the two cases PST and PHF are as follows:
where
where
The radiative heat flux in the xdirection is considered negligible in comparison to the ydirection. The radiative heat flux q_{r} is employed according to Rosseland approximation given by Siddheshwar and Mahabaleswar (2005) in the form
where σ* and k* are the StefanBoltzmann constant and the mean absorption coefficient, respectively.
Expanding T^{4} in a Taylor series about T_{∞}, we get
Neglecting terms beyond first degree in (TT_{∞}), we get
Substituting (12) in (10)
We use the following nondimensional influencing parameters in further steps.
Using expressions (7) to (9), (13) and the above nondimensional parameters,
Eq. 2 and 3, representing velocity and temperature
variation within the boundary layer can be reduced to the following Eq.
14 to 16.
The superscript ’ denotes differentiation with respect to η.
The boundary conditions (4) and (5) are rewritten in nondimensional form as follows:
Numerical procedure: The shooting method for the solution of nonlinear
differential equations basically involves choosing initial values for the concerned
derivatives in such a way that the end boundary conditions are satisfied with
in a prescribed numerical tolerance value. In this study, the numerical tolerance
value is chosen 10^{6} which is very close to zero for computational
purpose. The sequence of initial values is given by the secant method. The initial
value problem is solved using the fourth order RungeKutta scheme. The value
of η at ∞ i.e., η_{max} is so chosen that the solution
shows little further change for η larger than η_{max}. The
system of differential Eq. (14) to (16) together with the
boundary conditions (17) and (18) have been solved numerically using the RungeKutta
algorithm starting a systematic guessing of values for g”(0) f”’(0),
θ’(0) and g(0) and with the help of shooting technique until the boundary
conditions f’, f”, θ and g at η_{max} are satisfied
at the far end of the boundary. If the boundary conditions at η_{max}
are not satisfied, then the numerical technique uses a halfinterval method
to calculate the corrections to be applied to the initially estimated values
of f”(0), f”’(0), θ’(0) and g(0). This process is repeated
iteratively until the prescribed end values for f’, f”, θ and
g are obtained finally. The intermediate values at the end of each chosen interval
within the boundary are evaluated stage by stage.
RESULTS AND DISCUSSION
Figure 1 and 2 show the effects of viscoelastic
parameter (k_{1}) and permeability parameter (k_{2}) on the
flow velocity for two chosen values of k_{2} and k_{1}, respectively.

Fig. 1: 
Effect of the viscoelastic parameter k_{1}
on the velocity distribution (k_{2} = 1.0) 

Fig. 2: 
Effect of the permeability parameter k_{2}
on the velocity distribution (k_{1} = 0.2) 
The main effect of viscoelasticity is to gradually reduce the flow velocity
within the boundary layer, as may be seen from Fig. 1. It
is clear from Fig. 2 that the flow velocity also decreases
gradually with increase in porosity (k_{2}). It is observed however
that the influence of k_{1} is marginal as in Fig. 1
while the influence of k_{2} is relatively appreciable at low values
of k_{2} (<1).
The appropriate boundary conditions corresponding to PST and PHF cases are
used while solving the respective heat transfer Eq. (15)
and (16). The Rosseland approximation given by Siddheshwar
and Mahahbaleswar (2005) is used here to describe the radiative heat flux (q_{r})
in the energy equation.
Figure 36 show the results of the temperature
variation under varying values of k_{1}, k_{2}, N_{R}
and Pr. It may be seen from these figures that the effect of increasing k_{1}
and k_{2} is to increase the temperature at any point within the boundary.
The influences of k_{1} and k_{2} are, however, seen to be marginal
(Fig. 3 and 4).

Fig. 3: 
Effect of viscoelastic parameter k_{1}
on the temperature distribution. (a) PST case (b) PHF case 

Fig. 4: 
Effect of permeability parameter k_{2} on the temperature
distribution (a) PST case (b) PHF case 
However, lower values of k_{1} and k_{2} are desirable for
the fluid to act as an effective coolant in the application for extrusion polymer
process.
The effect of the radiation parameter (N_{R}) is found to be substantial as may be seen from Fig. 5. In contrast to this, influence of Prandtl number (Pr) is seen to be opposite to that of N_{R}. This means that, as Pr increases, the temperature decreases. This implies that a low value of N_{R} and a high value of Pr is a good combination for the fluid to perform as an effective cooling medium in applications such as extrusion processes in polymer industry.
Comparing the trend of results between PST and PHF cases for the same set of values of the influencing parameters, it is seen that at any chosen location within the boundary, the temperature is lower in PHF case than in PST case. This leads to the inference that PHF is a better option for the effective coolant action than PST case.

Fig. 5: 
Effect of the radiation number N_{R}
on the temperature distribution (a) PST case (b) PHF case 
Table 1: 
Values of wall temperature gradient θ’(0)
for the PST case and wall temperature g(0) for the PHF case for different
values of k_{1}, k_{2}, N_{R}, Pr and β 

Table 1 represents the values of the wall temperature gradient
θ^{1}(0) for the PST case and the wall temperature
g (0) for various values of k_{1}, k_{2}, N_{R}, Pr
and β. As all values of θ^{1}(0) are obtained negative for
the range of parameters studied herein, modulus values (θ’(0))
are tabulated.

Fig. 6: 
Effect of the Prandtl number on the temperature distribution
(a) PST case (b) PHF case 
This also implies that in this case, heat flows from the sheet to the fluid
region.
It may also be observed from the table that θ^{1}(0) decreases with increase in k_{1}, k_{2}, NR and β Further θ^{1}(0) increases with increase in Pr.
The wall temperature g(0) is seen to increase with increase in k_{1}, k_{2}, N_{R} and β but it decreases with increase in Pr.
In the absence of radiation, the results deduced from this study are seen to agree well with those obtained by Abel et al. (1998).
CONCLUSIONS
From this numerical study, the following conclusions are drawn:
• 
The velocity of the fluid decreases with increase in viscoelastic
parameter k_{1} and permeability parameter k_{2}. The influence
of k_{2} is however more marked than k_{1}. 
• 
The temperature at any point within the thermal boundary layer increases
with increase in k_{1}, k_{2} as also with increase in radiation
parameter N_{R}. The influence of Prandtl number Pr is however seen
to be opposite to the above. Further these results are seen to be valid
for both PST and PHF cases. 
• 
In applications where the fluid is to act as an effective coolant such
as in extrusion polymer processes, the PHF case may be preferred to PST
case since the former gives lower temperatures than in the latter case. 
• 
The wall temperature gradient θ^{1}(0) in the
PST case decreases with increase in k, k_{2}, N_{R} and
β but increases with increase in Pr. The opposite trend is seen for
the influence of these parameters on the wall temperature g(0) in the PHF
case. 
• 
The results obtained from this study for the particular case when the
effect of radiation is ignored (i.e., N_{R} = 0) agree well with
those obtained by Abel et al. (1998). 