Structures are exposed to damage during their service life which can severely affect their safety and functionality. Thus, it is important to monitor structures for the occurrence, location and extent of damage. Artificial neural networks (ANNs) as a numerical technique have been applied increasingly for damage identification with varied success. ANNs are inspired by human biological neurons and have been used to model some specific problems in many areas of engineering and science to achieve reasonable results. ANNs have the ability to learn from examples and then adapt to changing situations when sufficient input-output data are available. This paper presents the application of ANNs for detection of damage in a steel girder bridge using natural frequencies as dynamic parameters. Dynamic parameters are easy to implement for damage assessment and can be directly linked to the topology of structure. In this study, the required data for the ANNs in the form of natural frequencies will be obtained from experimental modal analysis. This paper also highlights the concept of ANNs followed by the detail presentation of the experimental modal analysis for natural frequencies extraction.
Key words: Artificial neural networks (ANNs), back propagation (BP), damage identification, natural frequency.
Copyright © 2018 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0