REDUCTION OF LINEAR DYNAMIC SYSTEMS IN FREQUENCY DOMAIN

Abstract

The development of reduced order models for the analysis and synthesis of control systems has been an area of active research during the last few decades. A wide variety of model order reduction methods have been proposed by the several authors in the area of model reduction. The model order reduction methods can be classified mainly 'into two categories : (i) time-domain order reduction methods; (ii) frequency-domain order reduction methods. The model order reduction methods applied to reduce the state-space models of the systems are called time-domain order reduction methods whereas the methods applied to the transfer function models of the systems are called frequency-domain order reduction methods. The main objective of model reduction is that the reduced order approximant should reproduce the significant characteristics of the parent system as closely as possible. The aim of this thesis is two fold : firstly to critically examine some of the existing model reduction methods and to develop some new model reduction methods for the reduction of linear time-invariant continuous time systems (CTS) in the frequency domain, and secondly to check their suitability for the design of controllers. The new developed methods for the reduction of continuous time systems are extended for the reduction of discrete time systems (DTS) in the frequency domain. The work included herein deals with the frequency domain model order reduction methods and control system design based on transfer function description of the system...