Please use this identifier to cite or link to this item: http://repositorio.unicamp.br/jspui/handle/REPOSIP/347105
Type: Artigo
Title: Estimation of beam material random field properties via sensitivity-based model updating using experimental frequency response functions
Author: Machado, M.R.
Adhikari, S.
Santos, J.M.C. Dos
Arruda, J.R.F.
Abstract: Structural parameter estimation is affected not only by measurement noise but also by unknown uncertainties which are present in the system. Deterministic structural model updating methods minimise the difference between experimentally measured data and computational prediction. Sensitivity-based methods are very efficient in solving structural model updating problems. Material and geometrical parameters of the structure such as Poisson’s ratio, Young’s modulus, mass density, modal damping, etc. are usually considered deterministic and homogeneous. In this paper, the distributed and non-homogeneous characteristics of these parameters are considered in the model updating. The parameters are taken as spatially correlated random fields and are expanded in a spectral Karhunen-Loève (KL) decomposition. Using the KL expansion, the spectral dynamic stiffness matrix of the beam is expanded as a series in terms of discretized parameters, which can be estimated using sensitivity-based model updating techniques. Numerical and experimental tests involving a beam with distributed bending rigidity and mass density are used to verify the proposed method. This extension of standard model updating procedures can enhance the dynamic description of structural dynamic models
Subject: Estimativa de parâmetro
Country: Países Baixos
Editor: Elsevier
Rights: Fechado
Identifier DOI: 10.1016/j.ymssp.2017.08.039
Address: https://www.sciencedirect.com/science/article/pii/S0888327017304727
Date Issue: 2018
Appears in Collections:FEM - Artigos e Outros Documentos

Files in This Item:
File Description SizeFormat 
000414113900011.pdf2.03 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.