A semi-localized reduced interface sampling scheme for dynamic substructuring method-based parametric reduced-order modelling
Citations

WEB OF SCIENCE

0
Citations

SCOPUS

0

초록

This paper proposes the semi-localized interface sampling based on conventional interpolation based parametric component mode synthesis (IB-PCMS). The conventional IB-PCMS addressed amount of computational resources required for the offline stage with sbstructural-level sampling, without considering full-system. Therefore, this approach enables more flexible reflection of design updates. On the other hand, the interface degrees-of freedom are fully retained without reduction in the CMS method for subcomponents synthesis. For this reason, the secondary eigenvalue analysis should be computed for interface reduction to achieve a highly-reduced system. In the previous IB-PCMS, substructures are synthesized and interface is reduced in the online stage. This additional computation hinders rapid responses with respect to parameter updating. To address this issue, the reduced interface sampling scheme is proposed by priory computing characteristic constraint (CC) modes. In addition, a semi-localized interface reduction method is newly proposed to efficiently construct reduced interface samples considering the parameter-dependent system. As a result, near real-time operation for the multi-query system can be achieved. © 2024 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

키워드

AviationDegrees Of Freedom (mechanics)Modal AnalysisStructural AnalysisComponent Mode SynthesisComputational ResourcesDynamic SubstructuringInterface ReductionLocalisedParametric ComponentsReduced Order ModellingReduced-order ModelSampling SchemesSampling-basedEigenvalues And Eigenfunctions
제목
A semi-localized reduced interface sampling scheme for dynamic substructuring method-based parametric reduced-order modelling
저자
Cheon, SeungheeLee, Jaehun
DOI
10.2514/6.2024-1437
발행일
2024-01
유형
Proceedings Paper
저널명
AIAA SciTech Forum and Exposition, 2024