World Applied Sciences Journal 17 (Special Issue of Applied Math): 10-15, 2012

ISSN 1818-4952
© IDOSI Publications, 2012

Applications of Runge-Kutta-Fehlberg Method and

Shooting Technique for Solving Classical Blasius Equation

Wan Mohd Khairy Adly Wan Zaimi, Biliana Bidin, Nor Ashikin Abu Bakar and Rohana Abdul Hamid

Institute of Engineering Mathematics, Universiti Malaysia Perlis, Malaysia

Abstract: The aim of this paper is to examine the classical boundary layer flow over a flat-plate namely
Blasius equation. The well-known Blasius equation is governed by the third order nonlinear ordinary
differential equation and then solved numerically using the Runge-Kutta-Fehlberg method with shooting
technique. The numerical results show a good agreement with the exact solution of Blasius equation and
consistent with prior published result. The accuracy of the proposed method is higher than other
approximation analytical solutions; hence suggest that proposed method is efficient and practical.

Key words: Blasius equation Runge-Kutta-Fehlberg method shooting technique

• •

INTRODUCTION equations. The idea of this technique is to assume the

solution as the summation of an infinite series. Later,
One of the equations arising in fluid mechanics is Aminikhah [6] add new dimension to this study by
the Blasius’s equation. This particular equation is a applying a new modification of homotopy perturbation
third order nonlinear differential equation contribute a method (NHPM) to solve numerically Blasius equation.
vital role in boundary layer theory in fluid mechanics. The growing concern in solving Blasius equation
The study of this famous equation has gained a great has motivated Wang [9] to examine this well-known
interest among researchers and mathematicians. To best equation by applying a new algorithm identified as
of author knowledge, the exact solution was first Adomian Decomposition Method (ADM). A few years
introduced by Blasius [1] using a series of expansion later, Hashim [10] has improved numerical solution
method. The first numerical solution for the Blasius reported by Wang [9] by using ADM -Padé approach
equation was given by Toepfer by employing Runge- and the results obtained has been compared to the
Kutta method. Later, Horwarth [2] produced more results using shooting method. Besides that, the Blasius
accurate solutions by improving Toepfer’s solution Mo equation also has been solved by Faiz and Albarakati
et al. [3]. It is worth mentioning that the numerical [11] using Wang’s transformations to determine the
result proposed by Horwarth [2] has been used by Mo approximate values of the second derivative at origin,
et al. [3], Cortell [4], Lin [5] and Aminikhah [6]. ƒ″(0). Afterwards, Mo et al. [3] presented numerical
Cortell [4] considered the third order nonlinear results for the discussed problem by using the method
differential equation, aƒ′″ +ƒƒ″=0 where a is a constant of weighted residuals.
and a prime denotes a differentiation with respect to η. There are various method can be applied in solving
For specific values a = 2, this equation has form of linear and nonlinear boundary value problem in
classical Blasius equation for the flat-plate problem. engineering and fluid mechanics such as Runge-Kutta
The numerical results were performed for several method, Homotopy Perturbation Method (HPM),
values of a in the interval 1≤a≤2 using Runge-Kutta Adomian Decomposition Method (ADM) and
algorithm for high-order initial value problems and Homotopy Analysis Method (HAM) [4, 7-9, 12].
consistent with Howarth [2]. On the other hand, Lin [5] However, in this present study, we solve the classical
has contribute numerical solutions of Blasius equation Blasius equation by employing combination of two
by using the parameter iteration method. Ganji et al. [7] numerical method namely Runge-Kutta-Fehlberg
have improved previous HPM results by Esmaeilpour method with shooting technique. The proposed methods
and Ganji [8] by employing Taylor series in solving have been used extensively by several researchers in
Blasius nonlinear differential equation. Ganji et al. [7] dealing with the problems of convective boundary layer
also reported that J. H. He in 1998 is the first who flows [13-15]. The solution procedure is much simpler
introduced the Homotopy Perturbation Method (HPM) and effective and the obtained result is of high
to solve the various linear and nonlinear differential accuracy.
Corresponding Author: Wan Mohd Khairy Adly Wan Zaimi, Institute of Engineering Mathematics, Universiti Malaysia Perlis,
Malaysia
11
 World Appl. Sci. J., 17 (Special Issue of Applied Math): 10-15, 2012

PROBLEM DEFINITION f ′ u,u

= =′ v (3)

The equation for the well-known Blasius equation

fv
is governed by the following nonlinear ordinary v′ + 0
=
2
differential equation (Blasius [1], Howarth [2]):

1 Here, ν = ƒ″. In accordance with conditions (2), we

f ′′′ ( η) + f (η ) f ′′ (η ) =0 (1) obtain
2

with boundary conditions ( 0) u=

f= ( 0) 0 , v=
(0 ) 0.3320573858

f= f=′ 0 at η = 0;
u = 1at η = η∞ (4)
f ′ → 1 as η→∞ (2)
We guess initial value of ν = ƒ″(0) =
where the “prime” denotes the derivatives with respect 0.3320573858 using Runge-Kutta-Felhberg method
to η. This equation is a form of the Blasius equation and then we solved the system of first order equations
relation for the flat plate [4]. Numerical results obtained Eq. (3) subject to conditions Eq. (4) using shooting
for the stream function ƒ(η), velocity profiles ƒ′(η) and technique. The code using Runge-Kutta-Felhberg
the variation of shear stress ƒ″(η). Equation (1) can be method in order to estimate initial value ƒ″(0) given as
writen as the first order system follow.

>
>

>

>

>

>

>

>

NUMERICAL RESULTS

The third order nonlinear diffential equation (1) with the boundary conditions (2) is solved
numerically by employing the fourth-fifth order Runge-Kutta-Felhberg method with shooting technique using
MAPLE software.
The numerical result is being compared with the published work by prior researchers and is presented in
Table 1-3 and also plotted in Fig. 1-4. Table 1-3 show the comparison of the present results by the Felhberg
fourth-fifth order Runge-Kutta method with shooting technique with Blasius [1], Howarth [2], method of
weighted residuals (MoWR) by Mo et al. [3], improved HPM by Ganji et al. [7], the HPM by Emaelpour
and Ganji [8] and Runge-Kutta algorithm for high-order initial value problems by Cortell [4] for the values ƒ (η),
ƒ′ (η) and ƒ″(η).
12
 World Appl. Sci. J., 17 (Special Issue of Applied Math): 10-15, 2012

Table 1: Comparison of the present results with published results for ƒ (η)
η Present method Ganji et al. [7] Esmaeilpour and Ganji [8] Blasius [1] Mo et al. [3] Howarth [2] Cortell [4]
0 0 0 0 0 0 0 0
0.2 0.00671 0.0066409 0.0069699
0.4 0.0267 0.0265598 0.0278758
0.6 0.0597 0.0597345 0.0626959
0.8 0.110 0.1061081 0.1113738
1.0 0.166 0.1655715 0.1738016 0.1656 0.1595 0.16557 0.16557
1.2 0.241 0.2379484 0.2498038
1.4 0.324 0.3229812 0.3391217
1.5 0.372 0.3701000 0.3701
1.6 0.425 0.4203202 0.4414008
1.8 0.530 0.5295172 0.5561797
2.0 0.652 0.6500224 0.6828833 0.6500 0.6459 0.65003 0.65003
2.2 0.782 0.7811878 0.8208206
2.4 0.925 0.9222734 0.9691873
2.5 1.000 0.9962000 0.9963
2.6 1.075 1.0724575 1.1270772
2.8 1.231 1.2308450 1.2935005
3.0 1.397 1.3964712 1.4674133 1.3968 1.3910 1.39682 1.39682
3.2 1.574 1.5682875 1.6477584
3.4 1.746 1.7451217 1.8335195
3.5 1.836 1.8350000 1.8377
3.6 1.933 1.9255896 2.0237911
3.8 2.112 2.1079372 2.2178650
4.0 2.298 2.2897787 2.4153361 2.3057 2.2882 2.30576 2.30576
5.0 3.283 3.2546 3.28329 3.28330
6.0 4.280 4.2456 4.27964 4.27965
7.0 5.279 5.2437 5.27926 5.27927
8.0 6.279 6.2433 6.27923 6.27925
9.0 7.279 7.2433 7.27923 7.27925

Table 2: Comparison of the present results with published results for ƒ′ (η)
η Present method Ganji et al. [7] Esmaeilpour and Ganji [8] Blasius [1] Howarth [2] Cortell [4]
0 0 0 0 0 0 0
0.2 0.0672 0.0664077 0.0696975
0.4 0.134 0.1327640 0.1393444
0.5 0.168 0.1658000 0.1659
0.6 0.201 0.1989370 0.2088105
0.8 0.269 0.2647088 0.2778800
1.0 0.330 0.3297796 0.3462538 0.3298 0.32979 0.32978
1.2 0.395 0.3937756 0.4135539
1.4 0.455 0.4562612 0.4793309
1.5 0.485 0.4867000 0.4868
1.6 0.515 0.5167558 0.5430747
1.8 0.575 0.5747553 0.6042289
2.0 0.630 0.6297563 0.6622097 0.6298 0.62977 0.62977
2.2 0.684 0.6812799 0.7164291
2.4 0.731 0.7288906 0.7663226
2.5 0.754 0.7511000 0.7513
2.6 0.776 0.7722037 0.8113803
2.8 0.812 0.8108693 0.8511819
3.0 0.848 0.8445213 0.8854328 0.8460 0.84605 0.84605
3.2 0.874 0.8726718 0.9140010
3.4 0.894 0.8945290 0.9369507
3.5 0.914 0.9027000 0.9130
3.6 0.919 0.9087097 0.9545718
3.8 0.935 0.9128113 0.9673977
4.0 0.955 0.9027999 0.9762106 0.9555 0.95552 0.95552
5.0 0.989 0.99155 0.99155
6.0 0.997 0.99898 0.99898
7.0 1.000 0.99992 0.99993
8.0 1.000 1.00000 1.00000
9.0 1.000 1.00000 1.00000

13
 World Appl. Sci. J., 17 (Special Issue of Applied Math): 10-15, 2012

Fig. 1: The variations of the function ƒ (η), ƒ′ (η) and

ƒ″(η) with η Fig. 3: Comparison of the present results with Blasius
[1] for ƒ′ (η)

Fig. 2: Comparison of the present results with Blasius

[1] for ƒ (η) Fig. 4: Comparison of the present results with Horwarth
[2] for ƒ″(η)
Table 3:Present results in comparison with published results for
ƒ″(η) CONCLUSION
η Present method Howarth [2] Cortell [4]
In this study, we have considered the classical
0 0.332057 0.33206 0.33206 Blasius problem. This nonlinear differential equation is
1.0 0.3250 0.32301 0.32301 successfully solved by employing Runge-Kutta-
2.0 0.2670 0.26675 0.26675 Fehlberg method with shooting method to obtain
3.0 0.1630 0.16136 0.16136 numerical solutions.
4.0 0.0672 0.06424 0.06423 It is found that present results are in good
5.0 0.1060 0.01591 0.01591 agreement compared to exact solutions by Blasius [1]
6.0 0.0020 0.00240 0.00240 for the values ƒ (η) and ƒ′ (η) as well as the values of
7.0 0.0000 0.00022 0.00022 ƒ″(η) in comparison with Howarth [2] as shown in
Table 1-3 and Fig. 2-4. The numerical results strongly
8.0 0.0000 0.00001 0.00001
display the efficiency and accuracy of the proposed
9.0 0.0000 0.00000 0.00000
method in solving the nonlinear equation.
14
 World Appl. Sci. J., 17 (Special Issue of Applied Math): 10-15, 2012

ACKNOWLEDGEMENT 8. Esmaeilpour, M. and D.D. Ganji, 2007.

Application of He’s homotopy perturbation
This work was supported by research grants from method to boundary layer flow and Convection
the Ministry of Higher Education, Malaysia (Project heat transfer over a flat plate. Physics Letters A.,
Code: FRGS/1/2011/SG/UNIMAP/03/7). 372 (1): 33-38.
9. Wang, L., 2004. A new algorithm for solving
REFERENCES classical Blasius Equation. Applied Mathematics
and Computation, 157 (1): 1-9.
1. Blasius, H. and Z.U. Math, 1908. Phys., 56: 1. 10. Hashim, I., 2006. Comments on “A new algorithm
2. Howarth, L., 1938. Notes. Rec. Roy. Soc. A, for solving classical Blasius equation” by L.
Vol: 164. Wang. Applied Mathematics and Computation,
3. Mo, L.F., S.H. You and J.X. Yin, 2008. 176 (2): 700-703
Application of the method of weighted residuals 11. Ahmad, F. and W.A. Albarakati, 2007. Application
to the Blasius equation. Journal of Physics: of Padé Approximation to solve the Blasius
Conference Series, Vol: 96. problem. Pror. Pakistan Amd. Sci., 44 (1): 17-19.
4. Cortell, R., 2005. Numerical solutions of the 12. Murat Ersen Berberler, Ahmet Yildirim and Syed
classical Blasius flat-plate problem. Applied Tauseef Mohyud-Din, 2011. Series Solutions
Mathematics and Computation, 170: 706-710 for Boundary Value Problems in Structural
5. Lin, J., 1999. A new approximate iteration solution Engineering and Fluid Mechanics. World Applied
of Blasius equation. Communications in Nonlinear Sciences Journal, 13 (12): 2428-2435.
Science and Numerical Simulation, 4 (2): 91-94. 13. Ishak, A., 2009. Mixed convection boundary layer
6. Aminikhah, H., 2010. An analytical approximation flow over a horizontal plate with thermal radiation.
for solving nonlinear Blasius equation by NHPM. Heat Mass Transfer, 46: 147-151.
Numerical Methods for Partial Differential 14. Deswita, L., A. Ishak and R. Nazar, 2010. Power-
Equations. 26 (6): 1291-1299. law Fluid Flow on a Moving Wall with Suction and
7. Ganji, D.D., H. Babazadeh, F. Noori, M.M Pirouz Injection Effects. Australian Journal of Basic and
and M. Janipour, 2009. An Application of Applied Sciences, 4 (8): 2250-2256.
Homotopy Perturbation Method for Non-linear 15. Singh, P., N.S. Tomer, S. Kumar and D. Sinha,
Blasius Equation to Boundary Layer Flow Over a 2010. MHD Oblique Stagnation-Point Flow
Flat Plate. International Journal of Nonlinear Towards a Stretching Sheet with Heat Transfer. Int.
Science, 7 (4): 399-404. J. of Appl. Math and Mech., 6 (13): 94-111.

15