![MathJax Logo](/templates/jsp/_style2/_tandf/pb2/images/math-jax.gif)
Abstract
In this paper, we propose an algorithm for constructing a pentadiagonal matrix with given prescribed three spectra. Sufficient conditions for solvability of the problem are given. We generate an algorithmic procedure to construct the solution matrices and we given a numerical example illustrating the construction algorithm.
Introduction
The free undamped infinitesimal vibrations, of frequency , of a thin straight beam of length
are governed by the Euler–Bernoulli equation
(1.1)
(1.1) Here
is Young’s modulus,
is the density,
is the cross sectional area at section
and
is the second moment of this area about the axis through the centroid at right angles to the plane of vibration. The problem can be discretized by assuming that the interval
is divided into
intervals with length
, and that
,
,
and
has the values
and
on the subintervals
, respectively. Figure shows a simple discrete model for the transversely vibrating Euler–Bernoulli beam. The parameters of discrete model can be defined as
(1.2)
(1.2) Figure shows the configuration around the mass
and the discrete system is obtained. For small displacement, the rotations are
The moment
needed to produce a relative rotation of the two rigid rods on either side of
is
Equilibrium of the rod linking
and
yields the shearing forces
Balance equation for mass
is
Here
and
denote external shearing forces and bending moments, respectively, applied to the ends. Suppose that the left hand end is clamped so that
The governing equations in the matrix form can be written as
Fig. 1 Discrete model of a vibrating beam.[Citation3]
![Fig. 1 Discrete model of a vibrating beam.[Citation3]](/cms/asset/deb18493-55ca-4b6f-b6ef-2b600c44ec34/gipe_a_797412_f0001_b.gif)
Fig. 2 The configuration around .[Citation3]
![Fig. 2 The configuration around mr.[Citation3]](/cms/asset/b565d6a9-da85-461b-b451-4f2c76ef8c97/gipe_a_797412_f0002_b.gif)
Statement of results
The matrices and
are of the form
(2.1)
(2.1) where
and
are pentadiagonal matrix of order
and
, respectively that
,
and
. Express the eigenvalue problem for
in terms of the normalized eigenvectors
and
of
and
, respectively. Thus if
is eigenvector of
then we can write
as follows
The eigenvalue problem
becomes
Premultiplying this equation by
and using orthogonality of matrices
and
and
we obtain
where
and
are eigenvalues of matrices
and
, in
respectively and
(2.2)
(2.2) Thus
so that
Theorem 2.1. If all
and
are distinct then
. Proof
Suppose that there exists an such that
is eigenvalue of
and
,
th component of eigenvector corresponding to
, is zero; thus
and
(2.3)
(2.3) Since
is eigenvector of
corresponding to
, thus
and from (2.3)
is eigenvalue of
which is a contradiction, since eigenvalues of matrices
and
are distinct.
By Theorem 2.1 we havewhere
and
are characteristic polynomials of the matrices
and
, respectively. This implies that
(2.4)
(2.4) where
Interlace condition (Equation1.14
(1.14)
(1.14) ) implies that
and
in (Equation2.4
(2.4)
(2.4) ) are positive. Now
and
may be computed from
(2.5)
(2.5) where
. From trace formula we can compute
and
as follows
(2.6)
(2.6) Now we consider eigenvalue problem for matrix
. Thus if
is eigenvector of
then we can write
as follows
(2.7)
(2.7) The eigenvalue problem becomes as follows
where
(2.8)
(2.8)
(2.9)
(2.9) Thus
so that for
we have
(2.10)
(2.10) Theorem 2.2. If all
and
are distinct then
. Proof
The proof is similar to Theorem 2.1. If we denote the coefficient matrix of the system (Equation2.10
(2.10)
(2.10) ) by
, then by Theorem 2.2,
for
. Thus
(2.11)
(2.11) where
denote characteristic polynomial of matrix
. Thus
(2.12)
(2.12) The system of nonlinear equations (Equation2.12
(2.12)
(2.12) ) can be solved by Newton’s method [Citation12] for
,
and
. With known
and
Equation (Equation2.9
(2.9)
(2.9) ) implies that
(2.13)
(2.13) where
. Since vectors
and
are orthonormal, multiplying the Equation (Equation2.8
(2.8)
(2.8) ) by
we find
(2.14)
(2.14) where
and
denote Euclidean norm of a vector. Also vectors
and
are orthonormal, multiplying the Equation (Equation2.2
(2.2)
(2.2) ) by
we obtain
(2.15)
(2.15) where
. In this procedure, a Gram-Schmidt process for finding orthonormal vectors
and elements
is necessary. Having
and
, matrices
and
in
can be computed by using the block Lanczos algorithm (see the Appendix 1). The following algorithm lists the steps for constructing the matrix
:
Construction Algorithm:
Compute
and
from (Equation2.6
(2.6)
(2.6) ).
Compute
and
from (Equation2.5
(2.5)
(2.5) ).
Compute
from (Equation2.13
(2.13)
(2.13) ) and
from (Equation2.14
(2.14)
(2.14) ).
Compute
from (Equation2.15
(2.15)
(2.15) ).
Construct matrices
and
from Lanczos algorithm (see Appendix 1).
Numerical examples
Here we illustrate the construction procedure by an example. We consider the Euler–Bernoulli beam equation as followsSuppose that
then the prescribed three spectra of discrete beam are as follows
These eigenvalues satisfy the conditions of Theorem 2.1, Theorem 2.2 and condition (Equation1.14
(1.14)
(1.14) ). Step 1 of algorithm gives
from (Equation2.4
(2.4)
(2.4) ) we obtain
and by step 2 we find
Solving the system (Equation2.12
(2.12)
(2.12) ) by using Newton’s method, we find
and step 3 gives
Using step 4, we find
Using the block Lanczos algorithm (Appendix 1), we find
Thus we constructed a symmetric pentadiagonal matrix
of the form (Equation1.12
(1.12)
(1.12) ) as follows
Computing the spectrum of the solution by MATLAB confirms the spectral data of the matrix
. With this
and by Appendix 2 we find masses
, stiffnesses
and lengths
of discrete beam as follows
The pentadiagonal matrices
in (Equation1.8
(1.8)
(1.8) ) and
in (Equation1.12
(1.12)
(1.12) ) are independent of
. But for different values of
we construct different discrete beam which have the same spectrum.
For we obtain
This pentadiagonal matrix have the same spectrum but for this we cannot construct a discrete beam since by (Equation1.13
(1.13)
(1.13) ) all
must be positive and all
must be negative.
Conclusion
In this paper, we formulated and solved the inverse problem of constructing a pentadiagonal matrix using three given spectra with interlacing property (Equation1.14
(1.14)
(1.14) ). This problem will have multiple solutions because in the constructing process we have different choices for
and vectors
and
. Also for pentadiagonal matrix we construct masses
, stiffnesses
and lengths
of corresponding discrete beam. Numerical example shows the accuracy of this method and details of the algorithm confirms the simplicity of the construction algorithm.
References
- Barcilon V. On the multiplicity of solutions of the inverse problem for a vibrating beam. Siam J. Appl. Math. 1979;37:605–613.
- Barcilon V. Inverse problems for a vibrating beam. J. Appl. Math. Phys. 1976;27:346–358.
- Gladwell GML. Inverse problems in vibration. New York (NY): Kluwer Academic; 2004.
- Singh KV. The transcendental eigenvalue problem and its application in system identification [Phd thesis]. Baton Rouge (LA): Department of Mechnical Engineering, Louisiana State University; 2003.
- Boley D, Golub GH. A survey of matrix inverse eigenvalue problems. Inverse Probl. 1987;3:595–622.
- Singh KV. Transcendental eigenvalue problem and its application. AIAA Journal. 2002;40:1402–1407.
- Chu MT, Diele F, Ragni S. On the inverse problem of constructing symmetric pentadiagonal Toeplitz matrices from their three largest eigenvalues. Inverse Probl. 2005;21:1879–1894.
- Caddemi S, Calia I. The influence of the axial force on the vibration of the Euler–Bernoulli beam with an arbitrary number of cracks. Arch. Appl. Mech. 2012;82:827–839.
- Caddemi S, Calio I. Exact closed-form solution for the vibration modes of the Euler–Bernoulli beam with multiple open cracks. J. Sound Vib. 2009;327:473–489.
- Gladwell GML. Minimal mass solutions to inverse eigenvalue problems. Inverse Probl. 2006;22:539–551.
- Ghanbari K, Mirzaei H, Gladwell GML. Reconstruction of a rod using one spectrum and minimal mass condition. Inverse Probl. Sci. Eng. 2013. doi:10.1080/17415977.2013.782543.
- Stoer J, Bulirsch R. Introduction to numerical analysis. New York (NY): Springer Verlag; 1993.
The block Lanczos algorithm [Citation3]
Suppose that and two vectors
and
are given. Let
, this algorithm constructs a pentadiagonal matrix
of the form
and orthogonal matrix
such that
(5.1)
(5.1) Here
for some integer
,
are symmetric matrices of order
and
are upper triangular matrix of order 2. The first
block of Equation (EquationAppendix 1.1
(1.1)
(1.1) ) gives
(5.2)
(5.2) since matrix
is orthogonal, premultiplying (EquationAppendix 1.2
(1.2)
(1.2) ) by
we obtain
(5.3)
(5.3) Equation (EquationAppendix 1.2
(1.2)
(1.2) ) can be written as
(5.4)
(5.4) Suppose
and
The first column of Equation (EquationAppendix 1.4
(1.4)
(1.4) ) gives
and the second column of (EquationAppendix 1.4
(1.4)
(1.4) ) gives
which leads to
so that
In this process, a Gram–Schmidt process for finding orthonormal vectors
and matrix
is necessary. By repeating the above process we can find matrices
and
. This algorithm can be applied for the
similarly. We have used this case in our numerical example.
Reconstruction of masses ![](//:0)
, stiffnesses ![](//:0)
and lengths ![](//:0)
[Citation3]
Suppose that we have computedwhere
and
is the pentadiagonal matrix in (Equation1.8
(1.8)
(1.8) ). We now untangle
to give masses
, stiffnesses
and lengths
.
First, we apply external static forces to masses 1 and 2, and deform the system as shown in Figure . For this configuration the static equation is
Fig. 3 Two static forces are needed to deflect all the masses by the same amount.[Citation3]
![Fig. 3 Two static forces are needed to deflect all the masses by the same amount.[Citation3]](/cms/asset/18e1fc18-c0a0-4a51-ade1-b8bd149c1f91/gipe_a_797412_f0003_b.gif)
Having found the masses , we find the lengths. We apply a single force
at
, and deform the system as shown in Figure . Now the equation
Fig. 4 One static force will deflect the beam as a straight line.[Citation3]
![Fig. 4 One static force will deflect the beam as a straight line.[Citation3]](/cms/asset/84561102-00a7-422a-a09d-05c1c4bdc21e/gipe_a_797412_f0004_b.gif)