31 March 2010 Support vector machine for abnormality detection on a cable-stayed bridge
Author Affiliations +
This paper applies support vector machine (SVM) to the field of structural health monitoring. SVM is a data processing technique that performs binary classification. The machine in its name indicates its association with machine learning, a category of algorithms that are able to solve classification problems by learning from example data given in a training process. This paper uses SVM for abnormality detection on data from a cable-stayed bridge's health monitoring system. The goal is to investigate whether the east end expansion joint is constraining the longitudinal motion of the bridge's main girder, which is suspected due to the results of a finite element updating procedure. Regarding the training process, distinct examples of the normal and abnormal expansion joint are unavailable from the health monitoring system. For this reason training examples are obtained from a finite element model. Accordingly, since SVM accuracy is highly dependent on the similarity between the training data and data being classified, the finite element modeling is a primary challenge of the paper's approach. The contributions of this paper include an application of SVM to an in-service structure, as well as a discussion on its performance and some limitations that affect its accuracy.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
David Vines-Cavanaugh, Yinghong Cao, Ming L. Wang, "Support vector machine for abnormality detection on a cable-stayed bridge", Proc. SPIE 7647, Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2010, 76471T (31 March 2010); doi: 10.1117/12.849248; https://doi.org/10.1117/12.849248

Back to Top