In this paper, a reduced basis approximation-based model order reduction for fast and reliable frequency sweep in the time-harmonic Maxwell's equations is detailed. Contrary to what one may expect by observing the frequency response of different microwave circuits, the electromagnetic field within these devices does not drastically vary as frequency changes in a band of interest. Thus, instead of using computationally inefficient, large dimension, numerical approximations such as finite element or boundary element methods for each frequency in the band, the point in here is to approximate the dynamics of the electromagnetic field itself as frequency changes. A much lower dimension, reduced basis approximation sorts this problem out. Not only rapid frequency evaluation of the reduced order model is carried out within this approach, but also special emphasis is placed on a fast determination of the error measure for each frequency in the band of interest. This certifies the accurate response of the reduced order model. The same scheme allows us, in an offline stage, to adaptively select the basis functions in the reduced basis approximation and automatically select the model order reduction process whenever a preestablished accuracy is required throughout the band of interest. Finally, real-life applications will illustrate the capabilities of this approach.

Accurate error estimation is crucial in model order reduction, both to obtain small reduced-order models and to certify their accuracy when deployed in downstream applications such as digital twins. In existing a posteriori error estimation approaches, knowledge about the time integration scheme is mandatory, e.g., the residual-based error estimators proposed for the reduced basis method. This poses a challenge when automatic ordinary differential equation solver libraries are used to perform the time integration. To address this, we present a data-enhanced approach for a posteriori error estimation. Our new formulation enables residual-based error estimators to be independent of any time integration method. To achieve this, we introduce a corrected reduced-order model which takes into account a data-driven closure term for improved accuracy. The closure term, subject to mild assumptions, is related to the local truncation error of the corresponding time integration scheme. We propose efficient computational schemes for approximating the closure term, at the cost of a modest amount of training data. Furthermore, the new error estimator is incorporated within a greedy process to obtain parametric reduced-order models. Numerical results on three different systems show the accuracy of the proposed error estimation approach and its ability to produce ROMs that generalize well.

No abstract

In this work, a model order reduction technique for fast frequency sweep in microwave filters and diplexers via the Reduced Basis Method is detailed. The Finite Element Method is used to solve the time-harmonic Maxwell's equations in these resonant circuit problems. Taking into account the electromagnetic field varies smoothly as a function of the frequency parameter, a low dimension system parameter manifold is identified. Thus, the original frequency-dependent Finite Element problem can be approximated by a model of reduced size using a Reduced Basis approximation. Based on the field frequency behavior in these resonant structures, the Reduced Basis space where the approximation holds is identified and characterized. Finally, special emphasis is placed in the behavior of the infsup constant used to provide an a posteriori error estimator, and its consequences in the Reduced Basis approximation are highlighted. Index Terms-Error analysis, finite element methods, Galerkin method, reduced basis methods, reduced order systems.

No abstract