Estimating relative eigenvalue errors in the Craig-Bampton method

Abstract

In this study, we propose an accurate error estimator for the Craig-Bampton (CB) method, which is a widely used component mode synthesis (CMS) method. The proposed error estimator can precisely predict relative eigenvalue errors in finite element models reduced by the CB method. To develop the error estimator, we propose an enhanced transformation matrix for the CB method and, using the transformation matrix, the error estimator is derived from the global (original) eigenvalue problem. In this paper, we demonstrate the robustness of the proposed error estimator through various numerical examples.