Share Email Print
cover

Proceedings Paper

Application of probabilistic neural network and static test data to the classification of bridge damage patterns
Author(s): Banfu Yan; Ayaho Miyamoto
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

This paper is primarily concerned with the applicaitonof static test data in conjunction withthe probabilistic neural nework (PNN) for the classificationof dmage patterns of a cable-stayed bridge. A total of 11 dmage patterns are considered by combinationof 5 typical dmage regions. Both training and testing data, derived from static analysis via finite element method (FEM), are contaminated with differnt noise level to simula eth eFe model and meausmeent erros. The study of damge pattern identificaitonis conducted by taking into account the change ratios of the deflection of the main beam and the towe runder loading as input neuron sof the pNN. The effects of noise levls, the types of damage patterns, and the number of input neurons on the identification accuracy ae investigated.Base don the classificaiton results some valuable conclusions were obtained.

Paper Details

Date Published: 18 August 2003
PDF: 12 pages
Proc. SPIE 5057, Smart Structures and Materials 2003: Smart Systems and Nondestructive Evaluation for Civil Infrastructures, (18 August 2003); doi: 10.1117/12.482684
Show Author Affiliations
Banfu Yan, Yamaguchi Univ. (Japan)
Ayaho Miyamoto, Yamaguchi Univ. (Japan)


Published in SPIE Proceedings Vol. 5057:
Smart Structures and Materials 2003: Smart Systems and Nondestructive Evaluation for Civil Infrastructures
Shih-Chi Liu, Editor(s)

© SPIE. Terms of Use
Back to Top