Identifier

etd-04142013-141514

Degree

Master of Science (MS)

Department

Electrical and Computer Engineering

Document Type

Thesis

Abstract

Structural health monitoring (SHM) is a very important field for many engineering disciplines. SHM deals with the monitoring of material structures periodically for assessing the lifetimes of the structures. There are various techniques for SHM. Non-destructive testing (NDT) is one of the most popular SHM tools to monitor structures. It demonstrates the indispensable advantage of providing structural health assessment without the need of intrusion. In this thesis, a new NDT tool for yield detection using ultrasonic signal processing is investigated. In this work, for the study of yield detection, steel specimen samples have been acquired, which were obtained from the laboratory of Department of Civil and Environmental engineering at Louisiana State University (LSU). An ultrasonic transducer then collected the signal data when these samples were tested. The data were preprocessed and segmented. For each acquired ultrasonic signal waveform, a total of three dominant echoes were extracted for the yield detection. A total of nine different signal features were extracted from these echoes for each ultrasonic signal. These nine features include time-domain features (signal amplitude, signal energy) and transform-domain features (wavelets, discrete Fourier transform, chirp Z-transform, discrete cosine transform, and discrete sine transform). Based on these aforementioned features, the linear discriminant analysis (LDA) technique is proposed to classify two situations (no-yield and yield). The proposed LDA-based classifier is compared with the conventional classifiers using individual features. The classifiers’ performances are evaluated using the receiver operating characteristics (ROC) plots. According to our experiments, it is discovered that the LDA-based classifier for yield detection is superior to all conventional classifiers using individual features, in terms of high detection rates subject to the fixed false detection rates.

Date

2013

Document Availability at the Time of Submission

Secure the entire work for patent and/or proprietary purposes for a period of one year. Student has submitted appropriate documentation which states: During this period the copyright owner also agrees not to exercise her/his ownership rights, including public use in works, without prior authorization from LSU. At the end of the one year period, either we or LSU may request an automatic extension for one additional year. At the end of the one year secure period (or its extension, if such is requested), the work will be released for access worldwide.

Committee Chair

Wu, Hsiao-Chun

Share

COinS