Order Reduction of Interval systems using Eigen Spectrum and Factor Division Algorithm

This paper describes a new approach for the reduction of higher order interval systems using a mixed method in which the numerator of the higher order model is reduced through factor division algorithm and reduced order denominator is obtained using Eigen Spectrum analysis by preserving the pole centroid and system stiffness of the original interval system. The stability of the reduced model is guaranteed, if the original high-order interval system is stable. The proposed method is compared with the other well known existing methods of order reduction. Further, the proposed method is substantiated by a numerical example.