MODEL REDUCTION AND CONTROL OF HIGHER ORDER SYSTEM VIA ITS REDUCED MODEL

Abstract

A large number of methods, for model reduction have been presented in both frequency and time domain in the literature. Three methods of model reduction using approximation, pade approximants using Routh array an stability margin technique in case of linear time invariant system are introduced in this dissertation. The stability margin technique is based on Optimization (minimization) technique in which error signal between original and reduced Order system is minimized subject to any critical frequency (gain cross over/phase cross over frequency) and also subjected to constraint derived from Routh table, which is governed by denominator of reduced model. This error function is minimized to evaluate the unknown coefficients of simplified/reduced model . Reduced model characteristics obtained by proposed technique, is more close to original system in both time and frequency domain. This method guarantees the stability of reduced model if original system is stable. A mixed method is ...