Premium
Guided waves for the health monitoring of sign support structures under varying environmental conditions
Author(s) -
Zhu Xuan,
Rizzo Piervincenzo
Publication year - 2013
Publication title -
structural control and health monitoring
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.587
H-Index - 62
eISSN - 1545-2263
pISSN - 1545-2255
DOI - 10.1002/stc.481
Subject(s) - structural health monitoring , ultrasonic sensor , guided wave testing , computer science , acoustics , truss , feature extraction , pattern recognition (psychology) , artificial intelligence , engineering , remote sensing , structural engineering , geology , physics
SUMMARY Guided ultrasonic waves are increasingly used in all those structural health monitoring applications that benefit from built‐in transduction, moderately large inspection ranges, and high sensitivity to small flaws. The wave‐based structural health monitoring of complex or thick structures can be difficult because of factors such as simultaneous propagation of many modes, large number of overlapping reflections, and mode conversion. Moreover, if the structures are subjected to environmental changes, variations on the characteristics of the detected signals are observed. This paper describes a health monitoring system for large trusses that combines the advantages of guided ultrasonic waves with the extraction of defect‐sensitive features aimed at performing a multivariate diagnosis of damage. The system was tested on a chord of a dismantled overhead sign support structure subjected to changing environmental conditions. Cold/warm temperatures and wet/snow boundary conditions were created. The system's hardware consisted of a data acquisition unit that controlled the generation and detection of ultrasonic signals by means of an array of piezoelectric transducers. The signals were processed using a feature‐based approach whereby features between two signals collected simultaneously or between a signal and a baseline were compared. The features were then fed into an unsupervised learning algorithm based on outlier analysis. Experimental results show that damage can be detected even in the presence of temperature and boundary condition variations. Copyright © 2011 John Wiley & Sons, Ltd.