Home // SENSORCOMM 2011, The Fifth International Conference on Sensor Technologies and Applications // View article
Authors:
Chunling Du
Jianqiang Mou
Landong Martua
Shudong Liu
Bingjin Chen
Guopeng Cao
Wen Xiang Yock
Jingliang Zhang
Frank L. Lewis
Keywords: active sensing; damage classification; feature extraction;finite element modeling; Lamb wave; principal component analysis; structure health monitoring; sensor network
Abstract:
In this work, the condition of a metallic structure is classified based on the acquired sensor data from a surfacemounted piezoelectric sensor/actuator network. The structure under consideration is an aluminum plate with riveted holes and possible crack damage in these holes is investigated. The sensor/actuator network uses diagnostic signals injected to piezoelectric actuators and received sensor signals to detect the crack. The damage classification system consists of three major components: sensitive signal acquisition, principal feature extraction and damage classification. An appropriate sine wave burst is used as diagnostic signals for actuators to transmit to sensors in order to detect the integrity of the structure. The combination of time-domain S0 waves from all sensitive sensor signals is directly used as features to detect damage. Since the time sequence of the extracted S0 waves is selected as the feature and has a high dimension, principal component estimation is applied to reduce the data dimension before entering the neural network training. Finally, in structure condition classification, a LVQ (learning vector quantization) neural network is used to classify structure conditions as healthy or damaged. In this paper, a number of FEM (finite element modeling) simulation results of sensor signals are taken as inputs to the neural network for training, since it is found that the FEM results have a good agreement with the experimental testing results on real plates. The performance of the classification is then validated by using these testing results.
Pages: 365 to 370
Copyright: Copyright (c) IARIA, 2011
Publication date: August 21, 2011
Published in: conference
ISSN: 2308-4405
ISBN: 978-1-61208-144-1
Location: Nice/Saint Laurent du Var, France
Dates: from August 21, 2011 to August 27, 2011