MINIMAL REALIZATION OF TRANSFER FUNCTION MATRICES AND ITS APPLICATION IN MODEL REDUCTION

Abstract

In this dissertation a method is describe a minimal realization of a linear time invariant multivariable system from a given rational transfer function matrices. State-space realization of dynamic equations is minimal, controllable and observable from this method. Minimal realization is done from Markov parameter and Taylor series which give the controllable and observable. Also describe application of minimal realization in model order reduction by Routh approximation, mixed method and aggregation method. First obtain the minimal realization of given transfer function using Markov parameter and Taylor series expansion, after that determine the transfer function from the matrices. Now reduces the transfer function order by Routh approximation and mixed method. Aggregation method reduces the matrices of require size, and then obtain the transfer function from reduced matrices. This transfer function is reduced order model transfer function. Also compare the reduced order model and original model by step response, and also compare by ISE.