Share Email Print

Proceedings Paper

Linear classification of system poles for structural damage detection using piezoelectric active sensors
Author(s): Jerome Peter Lynch
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The identification of damage in structural systems, including characterization of damage location and severity, is of extreme interest to the structural engineering profession. To date, many damage detection methods have been proposed that utilize global structural response measurements in the time and frequency domains to hypothesize the existence of structural damage. The accuracy and robustness of current damage detection methodologies could be improved through the use of active sensors. Active sensors, such as piezoelectric pads, impart low-energy acoustic excitations into structural elements and can record the corresponding system behavior. In this study, a novel methodology utilizing the input-output behavior of actively sensed structural elements is proposed. The poles of ARX time-series models describing modal frequencies and damping ratios are plotted upon the discrete-time complex plane and Perceptron linear classifiers employed to determine if poles of the structural element in an unknown state (damaged or undamaged) can be separated with those of the undamaged structure. If poles of the unknown state are separable from those of the undamaged state, the system is diagnosed as damaged. A simple cantilevered aluminum plate damaged by hack saw cuts is actively sensed by piezoelectric pads to show the efficacy of the proposed damage detection methodology. Furthermore, the number of misclassified poles and the final value of the Perceptron criterion function can be shown to be correlated to the severity of the damage.

Paper Details

Date Published: 29 July 2004
PDF: 12 pages
Proc. SPIE 5391, Smart Structures and Materials 2004: Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems, (29 July 2004); doi: 10.1117/12.539949
Show Author Affiliations
Jerome Peter Lynch, Univ. of Michigan (United States)

Published in SPIE Proceedings Vol. 5391:
Smart Structures and Materials 2004: Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems
Shih-Chi Liu, Editor(s)

© SPIE. Terms of Use
Back to Top