Damage classification using Adaboost machine learning for structural health monitoring

Daewon Kim; Michael Philen

doi:10.1117/12.882016

14 April 2011 Damage classification using Adaboost machine learning for structural health monitoring

Daewon Kim, Michael Philen

Proceedings Volume 7981, Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2011; 79812A (2011) https://doi.org/10.1117/12.882016
Event: SPIE Smart Structures and Materials + Nondestructive Evaluation and Health Monitoring, 2011, San Diego, California, United States

Abstract

In metallic structures, the first and second most frequent damages incurred are generally cracks and corrosions. Correct damage classification for these two damages is important since their phases can be developed with dissimilar patterns. In this research, damage classification using the Adaboost machine learning algorithm is investigated. To accomplish this, the physical differences of the two types of damages are defined and the most appropriate excitation signal is also determined. Various time-frequency methods are examined with the sensed damage signals to obtain a suitable signal processing method for damage classification. Among the methods examined, the spectrogram is chosen since it provides reliable results for these types of damages. With these results, the damage classification is performed through the Adaboost machine learning algorithm. The training samples for the algorithm are obtained from a finite element tool and experiments are also performed to get the testing samples. The analysis results show that correct damage classification is feasible using time-frequency representations and the Adaboost machine learning algorithm.

Citation Download Citation

Daewon Kim and Michael Philen "Damage classification using Adaboost machine learning for structural health monitoring", Proc. SPIE 7981, Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2011, 79812A (14 April 2011); https://doi.org/10.1117/12.882016

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available