Decoupling of second-order linear systems by isospectral transformation
- 33 Downloads
We consider the class of real second-order linear dynamical systems that admit real diagonal forms with the same eigenvalues and partial multiplicities. The nonzero leading coefficient is allowed to be singular, and the associated quadratic matrix polynomial is assumed to be regular. We present a method and algorithm for converting any such n-dimensional system into a set of n mutually independent second-, first-, and zeroth-order equations. The solutions of these two systems are related by a real, time-dependent, and nonlinear n-dimensional transformation. Explicit formulas for computing the \(2n \times 2n\) real and time-invariant equivalence transformation that enables this conversion are provided. This paper constitutes a complete solution to the problem of diagonalizing a second-order linear system while preserving its associated Jordan canonical form.
KeywordsSecond-order linear differential equations Quadratic matrix polynomials Diagonalization Isospectral systems
Mathematics Subject Classification15A22 34A30 70J10
RGS was supported by a Science Without Borders fellowship from the CAPES Foundation (grant no. 99999.011952/2013-00).
- 1.Meirovitch, L.: Principles and Techniques of Vibrations. Prentice Hall, Upper Saddle River (1997)Google Scholar
- 16.Lancaster, P., Psarrakos, P.: A note on weak and strong linearizations of regular matrix polynomials. Numerical analysis report 470, Manchester Centre for Computational Mathematics, Manchester (2005)Google Scholar