Frequency Limited Model Reduction Techniques With Error Bounds

The main issue with the pioneer frequency limited interval Gramian-based technique by Gawronski and Jaung is that it yields unstable reduced order models. There exist some techniques in literature that yield stable reduced order models, however, these techniques yield relatively more frequency response errors. In this brief, three frequency limited interval Gramians-based model order reduction techniques are proposed. The proposed techniques ensure the stability of the reduced order models, produce relatively lower frequency response error (as compared with existing stability preserving techniques) and yield the frequency response error bounds. Numerical examples along with comparison among different techniques are presented which show the effectiveness (relatively low approximation error as compared with other existing stability preserving model order reduction techniques) of the proposed techniques.