A reduction methodology using free-free component eigenmodes and Arnoldi enrichment

TL;DR

This paper introduces a novel model reduction method using free-free eigenmodes and Arnoldi enrichment, enabling smaller, more accurate, and updatable reduced models for modal analysis, outperforming traditional methods like Craig and Bampton.

Contribution

The paper develops a reduction methodology that produces smaller, updatable models independent of interface size, incorporating eigenmodes and Arnoldi enrichment for improved accuracy.

Findings

Achieves Craig and Bampton accuracy with smaller models

Enables partially updatable reduced models

Improves simulation efficiency in industrial contexts

Abstract

In order to perform faster simulations, the model reduction is nowadays used in industrial contexts to solve large and complex problems. However, the efficiency of such an approach is sometimes cut by the interface size of the reduced model and its reusability. In this article, we focus on the development of a reduction methodology for the build of modal analysis oriented and updatable reduced order model whose size is not linked to their contacting interface. In order to allow latter model readjusting, we impose the use of eigenmodes in the reduction basis. Eventually, the method introduced is coupled to an Arnoldi based enrichment algorithm in order to improve the accuracy of the reduced model produced. In the last section the proposed methodology is discussed and compared to the Craig and Bampton reduction method. During this comparison we observed that even when not enriched, our work enables us to recover the Craig and Bampton accuracy with partially updatable and smaller reduced order model.