Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/43854
Title: 1445. Structural dynamics-guided hierarchical neural-networks scheme for locating and quantifying damage in beam-type structures
Authors: Ding, ZC
Cao, MS
Jia, HL
Pan, LX
Xu, H
Keywords: Crack
Damage location
Damage quantification
Hierarchical neural networks
Modal frequency
Noise robustness
Steel beam
Issue Date: 2014
Publisher: Vibromechanika
Source: Journal of vibroengineering, 2014, v. 16, no. 7, p. 3595-3608 How to cite?
Journal: Journal of vibroengineering 
Abstract: Neural networks-based intelligent identification of structural damage mostly accentuates the capability of neural networks to recognize damage patterns but somewhat makes relatively less noticeable to the effect of structural dynamics in delivering an efficient damage identification scheme as well as damage features. At present, modal frequencies are often selected as inputs to neural networks for predicting location and/or severity of damage; however, this practice lacks solid support by structural dynamics. For that reason, this study investigates the use of structural dynamics-guided neural networks for damage identification. Structural dynamic analysis indicates that there are explicit relations between damage location and the ratio of changes of modal frequencies (RCMFs), and between damage severity and the change in modal frequencies (CMFs), before and after damage. These relations lay the foundation for creating a hierarchical neural-networks scheme to identify damage: using the RCMFs as inputs and damage location as the output to frame neural networks for locating damage; using the CMFs as inputs and damage severity as the output to establish neural networks for quantifying damage. This scheme features the guidance of structural dynamics on choosing damage indices and establishing a neural networks structure. The proposed scheme is numerically verified by identifying the location and severity of cracks in beams, with emphasis on noise robustness, and it is further experimentally validated using a set of steel beams with a through-width transverse crack, showing high accuracy and reliability in damage location and quantification.
URI: http://hdl.handle.net/10397/43854
ISSN: 1392-8716
Appears in Collections:Journal/Magazine Article

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

SCOPUSTM   
Citations

4
Last Week
1
Last month
Citations as of Nov 16, 2017

Page view(s)

75
Last Week
1
Last month
Checked on Nov 12, 2017

Google ScholarTM

Check



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