Model Reduction by New Clustering Method and Frequency Response Matching

Abstract

In this paper, mixed method of linear, time-invariant system model reduction method is suggested. A novel clustering algorithm based on Lehmer measure is utilized in the proposed method to obtain the reduced-order denominator polynomial. The selection of poles to form cluster center is based on the viewpoint of important poles contributing to the system is preserved by dominant pole algorithm. Having obtained the denominator polynomial of the reduced model, the coefficient of the numerator is found using the frequency response matching method. The reduction algorithm is fully computer oriented. The reduced model is stable if the original model is stable. Moreover, this method gives a good quality approximation in both the transient and the steady-state responses of the original system.