Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/76356
Title: System parameter identification from projection of inverse analysis
Authors: Liu, K
Law, SS 
Zhu, XQ
Keywords: Sensitivity analysis
Principal Component Analysis
Model updating
Noise
Projection
Parameter estimation
Issue Date: 2017
Publisher: Academic Press
Source: Journal of sound and vibration, 2017, v. 396, p. 83-107 How to cite?
Journal: Journal of sound and vibration 
Abstract: The output of a system due to a change of its parameters is often approximated with the sensitivity matrix from the first order Taylor series. The system output can be measured in practice, but the perturbation in the system parameters is usually not available. Inverse sensitivity analysis can be adopted to estimate the unknown system parameter perturbation from the difference between the observation output data and corresponding analytical output data calculated from the original system model. The inverse sensitivity analysis is re-visited in this paper with improvements based on the Principal Component Analysis on the analytical data calculated from the known system model. The identification equation is projected into a subspace of principal components of the system output, and the sensitivity of the inverse analysis is improved with an iterative model updating procedure. The proposed method is numerical validated with a planar truss structure and dynamic experiments with a seven-storey planar steel frame. Results show that it is robust to measurement noise, and the location and extent of stiffness perturbation can be identified with better accuracy compared with the conventional response sensitivity-based method.
URI: http://hdl.handle.net/10397/76356
ISSN: 0022-460X
EISSN: 1095-8568
DOI: 10.1016/j.jsv.2017.02.042
Appears in Collections:Journal/Magazine Article

Access
View full-text via PolyU eLinks SFX Query
Show full item record

SCOPUSTM   
Citations

1
Citations as of Dec 13, 2018

WEB OF SCIENCETM
Citations

1
Last Week
0
Last month
Citations as of Dec 13, 2018

Page view(s)

23
Last Week
2
Last month
Citations as of Dec 16, 2018

Google ScholarTM

Check

Altmetric


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