Asymptotic Method of Krylov-Bogoliubov-Mitropolskii for Fifth Order Critically Damped Nonlinear Systems
Md. Nazrul Islam1, Md. Mahafujur Rahaman2, *, M. Abul Kawser1
1Department of Mathematics, Islamic University, Kushtia, Bangladesh
2Department of Computer Science & Engineering, Z. H. Sikder University of Science & Technology, Shariatpur, Bangladesh
Email address:
To cite this article:
Md. Nazrul Islam, Md. Mahafujur Rahaman, M. Abul Kawser. Asymptotic Method of Krylov-Bogoliubov-Mitropolskii for Fifth Order Critically Damped Nonlinear Systems. Applied and Computational Mathematics. Vol. 4, No. 6, 2015, pp. 387-395. doi: 10.11648/j.acm.20150406.11
Abstract: In oscillatory problems, the method of Krylov–Bogoliubov–Mitropolskii (KBM) is one of the most used techniques to obtain analytical approximate solution of nonlinear systems with a small non-linearity. This article modifies the KBM method to examine the solutions of fifth order critically damped nonlinear systems with four pairwise equal eigenvalues and one distinct eigenvalue, in which the latter eigenvalue is much larger than the former four pairwise eigenvalues. This paper suggests that the results obtained in this study correspond accurately to the numerical solutions obtained by the fourth order Runge-Kutta method. This paper, therefore, concludes that the modified KBM method provides highly accurate results, which can be applied for different kinds of nonlinear differential systems.
Keywords: KBM, Asymptotic Method, Critically Damped System, Nonlinearity, Runge-Kutta Method, Eigenvalues
1. Introduction
In oscillatory problems, the method of Krylov–Bogoliubov–Mitropolskii (KBM) [1,2] is particularly convenient, and is the vastly used technique to obtain analytical approximate solution of nonlinear systems with a small non-linearity. The method was, in fact, developed by Krylov and Bogoliubov [2] for obtaining periodic solutions, which was amplified and justified by Bogoliubov and Mitropolskii [1], and later extended by Popov [3] and Meldelson [4] for damped nonlinear oscillations. Murty [5] developed a unified KBM method for solving second-order nonlinear systems. Sattar [6] studied a third-order over-damped nonlinear system. Bojadziev [7] studied the damped oscillations modeled by a three dimensional nonlinear system. Shamsul and Sattar [8] developed a method for third order critically-damped nonlinear equations. Rokibul and Akbar [9] investigated a new solution of third order more critically damped nonlinear systems. Shamsul and Sattar [10] presented a unified KBM method for solving third-order nonlinear systems. Akbar et al. [11] presented a method for solving the fourth-order over-damped nonlinear systems. Rokibul et al. [12] presented a new technique for fourth order critically damped nonlinear systems with some conditions. Rahaman and Rahman [13] suggested analytical approximate solutions of fifth order more critically damped systems in the case of smaller triply repeated roots. Rahaman and Kawser [14] also proposed asymptotic solutions of fifth order critically damped nonlinear systems with pairwise equal eigenvalues and another is distinct.
The aim of this article is to obtain the analytical approximate solutions of fifth order critically damped nonlinear systems by extending the KBM method. In this study, it is suggested that the results obtained by the perturbation solution have been compared with those obtained by the fourth order Runge–Kutta method.
2. The Method
We are going to propose a perturbation technique to solve fifth order non-linear differential systems of the form
(1)
where and
stand for the fifth and fourth derivatives respectively, and over dots are used for the first, second and third derivatives of x with respect to t;
are constants,
is a sufficiently small parameter and
is the given nonlinear function. As the unperturbed equation (1) has five real negative eigenvalues, where four eigenvalues are pairwise equal and the other one is distinct. Here, the distinct eigenvalue is much larger than the pairwise equal eigenvalues. Now, suppose that the eigenvalues are
When ε = 0, the equation (1) becomes linear and the solution of the corresponding linear equation becomes
(2)
where ,
,
,
and
are constants of integration.
When ε ≠ 0 following Shamsul [15], an asymptotic solution of the equation (1) is sought in the form
(3)
where a, b, c, d and h are the functions of t and they satisfy the first order differential equations
(4)
Now differentiating (3) five times with respect to t, substituting the value of x and the derivatives in the original equation (1) utilizing the relations presented in (4) and finally extracting the coefficients of ε, we obtain
(5)
Where
And
We have expanded the function in the Taylor’s series (Sattar [16], Shamsul [17, 18], Shamsul and Sattar [8]) about the origin in power of t. Therefore, we obtain
(6)
Thus, using (6), the equation (5) becomes
(7)
Following the KBM method, Murty and Deekshatulu [19], Sattar [16], Shamsul [18], and Shamsul and Sattar [8, 20] imposed the condition that does not contain the fundamental terms (the solution (2) is called the generating solution and its terms are called the fundamental terms) of
. Therefore, equation (7) can be separated for unknown functions
and
in the following way:
(8)
(9)
Now equating the coefficients of from equation (8), we obtain
(10)
(11)
Here, we have only two equations (10) and (11) for determining the unknown functions A1, B1, C1 , D1 and H1 Thus, to obtain the unknown functions A1, B1, C1, D1 and H1, we need to impose some conditions (Shamsul [18,20,21,23]) between the eigenvalues. Different authors have imposed different conditions according to the behavior of the systems, such as Shamsul [21] imposed the condition
In this study, we have investigated solutions for the cases and
. Therefore, we shall be able to separate the equation (11) for unknown functions B1 and D1; and solving them for B1 and D1 substituting the values of B1 and D1 into the equation (11) and applying the conditions
and
, we can separate the equation (12) for three unknown functions A1, C1 and H1; and solving them for A1, C1 and H1. Since
and
are proportional to small parameter, they are slowly varying functions of time t, and for first approximate solution, we may consider them as constants in the right side. This assumption was first made by Murty and Deekshatulu [19]. Thus, the solutions of the equation (4) become
(12)
Equation (9) is a non-homogeneous linear ordinary differential equation; therefore, it can be solved by the well-known operator method. Substituting the values of a, b, c, d, h and in the equations (3), we shall get the complete solution of (1). Therefore, the determination of the first approximate solution is complete.
3. Example
As an example of the above method, we have considered the Duffing type equation of fifth order nonlinear differential system:
(13)
Comparing equation (13) and equation (1), we obtain
Therefore,
(14)
Now comparing equations (6) and (14), we obtain
(15)
For equation (13), the equations (9) to (11) respectively become
(16)
(17)
(18)
Since the relations and
among the eigenvalues, then the equation (18) can be separated for the unknown functions
and
in the following way:
(19)
(20)
Solving equations (19) and (20), we get
(21)
(22)
Where
Using the values of and
in equation (17), we obtain
(23)
Again, applying the conditions and
in equation (23), we obtain the following equations for unknown functions
and
:
(24)
(25)
(26)
Solving equations (24), (25) and (26), we obtain
(27)
(28)
(29)
The solution of the equation (3.3.4) for is
(30)
where
Substituting the values of and
from equations (27), (21), (28), (22) and (29) into equation (4), we obtain
(31)
Here, all of the equations (31) have no exact solutions. However, sinceand
are proportional to the small parameter
, they are slowly varying functions of time t. Consequently, it is possible to replace a, b, c, d and h by their respective values obtained in linear case (i.e., the values of a, b, c, d and h obtained when
= 0) in the right hand side of equations (31). This type of replacement was first introduced by Murty and Deekshatulu [24], and Mutry et.al. [19] to solve similar types of nonlinear equations. Therefore, the solutions of equation (31) are
(32)
Hence, we obtain the first approximate solution of the equation (13) as:
(33)
where a, b, c, d and h are given by the equations (32) and is given by (30).
4. Results and Discussion
The perturbation solution is usually compared to the numerical solution to test the accuracy of the approximate solution obtained by a certain perturbation method. Therefore, we have first considered the eigenvalues
and
We have computed
using (33), in which
and
are obtained from (32) and
is calculated from equation (30) together with initial conditions
and
when
. The result obtained from (33) for various values of t, and the corresponding numerical solution obtained by a fourth order Runge-Kutta method is plotted in the Fig. 1.
Fig. 1. Perturbation results are plotted by continuous line and numerical results are plotted by dotted line.
Again, we have computed from (33) by considering values of
and
We have computed
using (33), in which
and
are obtained from (32) and
is calculated from equation (30) together with initial conditions
and
when
The result obtained from (33) for various values of t, and the corresponding numerical solution obtained by a fourth order Runge-Kutta method is plotted in the Fig. 2.
Fig. 2. Perturbation results are plotted by continuous line and numerical results are plotted by dotted line.
Finally, we have computed from (33) by considering values of
and
We have computed
using (33), in which
and
are obtained from (32) and
is calculated from equation (30) together with initial conditions
and
when
The result obtained from (33) for various values of t, and the corresponding numerical solution obtained by a fourth order Runge-Kutta method is plotted in the Fig. 3.
Fig. 3. Perturbation results are plotted by continuous line and numerical results are plotted by dotted line.
5. Conclusion
In conclusion, it is suggested that, in this article, the KBM method has been modified and applied successfully to the fifth order more critically damped nonlinear systems. In relation to the fifth order critically damped systems, the solutions are obtained in such circumstances where the four eigenvalues are pairwise equal and another eigenvalue is distinct. Normally, in the KBM method, it is noticed that much error occurs in the case of rapid changes of x with respect to time t. However, it is suggested that all the aforementioned results obtained in this paper correspond accurately to the numerical solutions obtained by the fourth order Runge-Kutta method. It is, therefore, concluded that the modified KBM method provides highly accurate results, which can be applied for different kinds of nonlinear differential systems.
Acknowledgement
The authors are grateful to Mr. Md. Mizanur Rahman, Associate Professor, Department of Mathematics, Islamic University, Bangladesh, for his invaluable comments on the early draft of this paper. The authors are also thankful to Mr. Md. Imamunur Rahman for his assistance in editing this paper.
References