MODEL REDUCTION IN SCHUR BASIC WITH POLE RETENTION an‎d H2 NORM ERROR BOUND

Model order reduction is a field of study to identify low-order systems to replace the high-order systems while preserving the essential properties of the original system. One of critical properties of the high-order systems that need to be preserved in the order reduction is dominant poles. This paper introduces a model order reduction algorithm by preserving dominant poles of the original system in the reduced-order system. By evaluating the dominance of the poles according to the H2 standard, the algorithm will convert the matrix-A of the original system to an upper triangular form and then arrange the poles of the system according to the decreasing importance on the main diagonal of the upper triangular matrix. Using the truncation technique, the algorithm will retain the dominant poles corresponding to the top positions on the main diagonal of the upper triangular matrix to obtain the reduced-order system. Applying the algorithm to reduce the order of the high-order controller and the high-order filter shows that the algorithm is able to arrange the poles according to the H2 dominance and preserve the dominant poles of the original system in the order reduction system.