In recent years, model order reduction (MOR) has been interested in more and more scientists. A lot of MOR algorithms have been introduced by many different approaches, among which preserving the dominant poles of the original system and Hankel singular values of the original system in order reduction system are appropriate approaches with many advantages. The article introduces a new MOR algorithm applied for stable and unstable linear systems, based on the idea of preserving the dominant poles of the original system during the order reduction. The algorithm will switch matrix-A of the original high-order system into the upper triangular matrix, then arrange the poles under the measure of dominance- H, H₂, and mixed points on the main diagonal of upper triangular matrix-A, in order to attain a small error order reduction and preserve dominant poles simultaneously. The effectiveness of the new algorithm is illustrated through the order reduction of the high-order controller. Simulation results have proven the correctness of the algorithm.

近年来，模型阶数减少（MOR）引起了越来越多科学家的兴趣。许多不同的方法引入了许多MOR算法，其中保留原始系统的主导极点和降阶系统中原始系统的Hankel奇异值是具有许多优点的合适方法。本文介绍了一种新的MOR算法，该算法适用于稳定和不稳定的线性系统，其基本思想是在降阶过程中保留原始系统的主导极点。该算法将原始高阶系统的矩阵A转换为上三角矩阵，然后以优势度H，H ₂排列极点，以及上三角矩阵A的主对角线上的混合点，以便获得较小的误差阶数减少并同时保留主导极点。通过高阶控制器的降阶来说明新算法的有效性。仿真结果证明了该算法的正确性。