Order Reduction of LTI Systems and Their Qualitative Comparison

Abstract

ABSTRACTA novel model order-reduction method is proposed for both single and multivariable systems; the method is based on preserving dominant frequency selection. The main aim of the paper is complexity reduction and obtains a suitable lower order system from original higher order model. The proposed method preserves the dominant frequencies of the higher order model in the reduced denominator polynomial and, finally, the lower order model is obtained by indirectly matching the initial time moments of original and reduced systems. The suitability of the proposed method is judged by comparing the step responses of the higher and lower order systems. The performances of the systems are measured in terms of integral square error (ISE). It is seen from the numerical examples that the results are quite competitive as compared to other methods existing in the literature, thus justifying the applicability of the proposed method. The method is also extended to order reduction of discrete time systems.