Decomposition of Time-Scales in Linear Systems Using Dominant Eigenspace Power Iterations and Matched Asymptotic Expansions.
Abstract
Many control theory concepts are valid for any system order, however, their actual use is limited to low order models. Large scale systems result not only in a formidable amount of computation, but also in ill-conditioned initial and two point boundary value problems. The interaction of fast and slow phenomena in high-order systems results in stiff numerical problems which require expensive integration routines. The singular perturbations approach to decomposing fast ad slow phenomena involves using a time-scale separation technique. In ths case a reduced order 'steady state' and 'boundary layer' solutions are obtained from a high order model. Control designs and simulations for the high order model are then carried out on the reduced order subsystems. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1979
- Accession Number
- ADA085260
Entities
People
- P. V. Kokotovic
- R. G. Phillips
Organizations
- University of Illinois Urbana–Champaign