Repeated Eigenvalues/Decoupled Systems

I am computing the eigenvalues and eigenvectors for a mechanical
system that degenerates to multiple, unconnected systems under some
conditions. When it decouples, I get repeated eigenvalues at the
expected frequencies. However, I'm having problems figuring out how
to get the eigenvectors to decouple.

If I take the eigenvectors returned by [V,D]=eig(sys), for the
degenerate modes, they involve contributions from both subsystems
(probably not too surprising).

If I compute the eigenvectors using null as in

Vn = null(sys - Dn*eye(size(sys)),'r')

they decouple nicely - at least so far.

It seems like there should be an reasonable way to work with the
eigenvectors from eig, and the known structure of the system to get a
decoupled set. Any suggestions?

Thanks, Erik
.

Relevant Pages

  • Re: Repeated Eigenvalues/Decoupled Systems
    ... I am computing the eigenvalues and eigenvectors for a mechanical ... degenerate modes, they involve contributions from both subsystems ... they decouple nicely - at least so far. ...
    (comp.soft-sys.matlab)
  • Re: Is this matrix diagonalizable?
    ... So I found that we have two repeated eigenvalues, namely lambda = 2. ... Therefore among the 4 eigenvectors there are two eigenvectors which ... indep0endent and thus will conmform a basis of the vector space ==> ...
    (sci.math)
  • Re: Is this matrix diagonalizable?
    ... So I found that we have two repeated eigenvalues, namely lambda = 2. ... Therefore among the 4 eigenvectors there are two eigenvectors which ... indep0endent and thus will conmform a basis of the vector space ==> ...
    (sci.math)
  • Is this matrix diagonalizable?
    ... So I found that we have two repeated eigenvalues, namely lambda = 2. ... Therefore among the 4 eigenvectors there are two eigenvectors which ... be linearly independent because a vector is a multiple of itself. ...
    (sci.math)