Untitled

Mahanfar, A.; Bila, Stéphane; Aubourg, Michel; Verdeyme, Serge; Boichon, Christelle; Puech, J.; Lapierre, L.

doi:10.1002/mmce.v19:2

Cited by 1 publication

(1 citation statement)

References 201 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Through an automatic training process, ANN can learn the relationship between EM responses and geometrical parameters. The trained model provides accurate and fast prediction for the EM behavior of microwave components with geometrical parameters as variables and can be subsequently implemented in circuit and system designs [13,14]. However, how to develop multi-physics parametric models using ANN techniques remains an open topic.…”

Section: Introductionmentioning

confidence: 99%

Multi-Physics Parametric Modeling of Microwave Passive Components Using Artificial Neural Networks

Yan¹,

Zhang²,

Jin³

et al. 2018

PIER M

View full text Add to dashboard Cite

In this paper, a novel multi-physics parametric modeling approach using artificial neural networks (ANNs) for microwave passive components is proposed. In the proposed approach, the ANN is used to learn the nonlinear relationships between electromagnetic (EM) behaviors and multiphysics design variables. The trained model can accurately represent the EM responses of the passive components with respect to the multi-physics input parameters. Therefore, the proposed model can provide accurate and fast prediction of EM responses using low computational cost and little time for multi-physics design. The advantage of the proposed model is demonstrated by two microwave examples: the proposed model can save about 98% computational cost compared with the EM model, and the CPU time of the proposed model is less than 0.1 s while that of the EM model needs many minutes.

show abstract

Section: Introductionmentioning

confidence: 99%