Response statistics prediction for structures with uncertain attachments using a Krylov approach

Abstract

This paper presents an approach to calculate the response statistics of a structure comprising a main, host structure with an attachment whose coupling location is not exactly known. The proposed approach is based on a Krylov model order reduction method, which reduces the order of a finite element model of the structure enabling an efficient way to estimate the statistics of the structural response. The multi-point constraint method is used to connect the finite element models of the attachment and the structure. One (or more) analysis based on Chebyshev nodes is performed for the attachment located at a given position(s) to obtain the local projection matrices. Afterwards, a Parametric Model Order Reduction method is applied to reduce the model order. Uncertainty in the location of the attachment is included in a Monte Carlo simulation. The method is illustrated using two numerical examples to demonstrate its accuracy and efficiency.