Abstract
This article considers the problem of instability in similarity transformation-based model reduction methods. Given a stable model G(s) of order n, a general framework is proposed that parameterises a large set of stable reduced models of order k n that each can be obtained by truncating a realisation of G(s). As an application of this result it is shown how different reduction methods can be modified, if they fail to maintain stability.
Original language | English |
---|---|
Pages (from-to) | 297-307 |
Number of pages | 11 |
Journal | International Journal of Systems Science |
Volume | 40 |
Issue number | 3 |
DOIs | |
State | Published - Mar 2009 |
Keywords
- Lyapunov equation
- Model reduction
- Stability