Sit Back, Relax and Model Order Reduction for Microwave Filter Design
Abstract
The increasing deployment of communication services is urging microwave filter industry to satisfy stringent electrical specifications in microwave devices and, in addition, in rather short delivery times. These electromagnetic devices require time-consuming full-wave simulations, by means of commercial software. However, there is plenty of room for improvement in this strategy to satisfy industrial needs. Instead of using engineering brute force, which can be expensive and computer-demanding, we can use the so-called model order reduction techniques.A reduced-order model is a mathematically equivalent model to a more complicated full-order model, still maintaining the same behavior of the original one within a specific parameter range, but in a much simpler setting.This work introduces what model order reduction can do for you in microwave filter full-wave design. Traditionally, fast frequency sweeps have been provided by these reduced-order models. However, there are plenty of more options available. The electromagnetic system matrix is compressed to its minimum size, i.e., the order of the microwave filter plus a few more, say, three or four more. This reduced-size system matrix captures all the physics in the 3D model of your electromagnetic device. As a result, a more in-depth zero-pole diagram can be obtained with ease, where the dominant electromagnetic poles, reflection and transmission zeros are clearly identified. The experienced eyes of the microwave engineer no longer need to guess where these poles and zeros are located by looking at the S-parameter response. This model order reduction approach detail their position on the complex plane, in the Laplace domain.In addition, this reduced-order model, i.e., the reduced-size system matrix, can be arranged to identify the electromagnetic couplings among the resonators in the microwave device under study. This allows the microwave engineer to know which couplings are actually present in the full-wave simulation of the electromagnetic device, including the leakage ones. As a result, the tuning process of the device turns into a simple task, adjusting only the right couplings to target the electrical specifications.Several examples will show the capabilities and possibilities of model order reduction approaches in microwave filter design.
Fields of Study & Subjects
Model Reduction and Neural Networks
Microwave Engineering and Waveguides
Electromagnetic Simulation and Numerical Methods
Publication details
- Journal / source
- IEEE International Microwave Filter Workshop (IMFW)
- Pages
- 1–3
- Publication date
- 7 February 2026
- Language
- EN