|
Descripción
|
|
|---|---|
| In this work, we detail a domain decomposition scheme for design purposes. When developing a design by means of an optimization technique, the analysis domain is typically modi?ed, e.g., its shape, the electromagnetic properties of the materials or their shapes, until the required speci?- cations are satis?ed. It seems clear that the global structure ought to be analyzed each time a modi?cation is carried out. However, there is no point in having to solve the whole electromagnetic problem while only a small portion of domain has changed. What we propose is to isolate those parts from the remaining analysis domain in such a way that every modi?cation carried out on these be transparent to the remaining analysis domain, which should only be analyzed once, and thus concentrate on the analysis of those parts that evolve. In other words, we aim to use our com-puting resources during the optimization process only in those parts of the analysis domain that change. In order to carry out this strategy, a direct domain decomposition approach in the FEM is considered, and thus, the subdomains are electromagnetically described by full-wave matrix-valued transfer functions, namely, generalized scattering matrices. We pay special attention to come up with a formulation dropping singularities in frequency out of the Maxwell operator. Once these transfer functions are determined, the response of the overall system is recovered by the scattering matrix connection. This methodology is compatible with a reduced-order model approach and, as a result, a reliable fast frequency sweep technique is carried out. | |
|
Internacional
|
Si |
|
Nombre congreso
|
5th International Workshop on Microwave Filters |
|
Tipo de participación
|
960 |
|
Lugar del congreso
|
Tolouse |
|
Revisores
|
Si |
|
ISBN o ISSN
|
|
|
DOI
|
|
|
Fecha inicio congreso
|
01/10/2012 |
|
Fecha fin congreso
|
03/10/2012 |
|
Desde la página
|
1 |
|
Hasta la página
|
1 |
|
Título de las actas
|
Proc. of the 5th International Workshop on Microwave Filters |