Translator Disclaimer
17 April 2013 Abnormal behavior detection in the Jeremiah Morrow Bridge based on the long term measurement data patterns
Author Affiliations +
Calibration of a finite element model based on measurement-data for complex structures is usually costly and sometimes not applicable. In this article, a methodology for detecting abnormal behavior including slow aging degradations of a structure solely based on historical patterns of the measurement data will be introduced. In the first step, principal components of the truck load test measurement data - that is centered and scaled - are calculated. In the second step, unsupervised classification is applied to the score data that is regenerated based on the major principal components. The same algorithm is applied to the measurement data of the bridge response to the sharp temperature change as well. Finally, the specified algorithm is applied separately to the collected static data from the Jeremiah Morrow Bridge (more than four years) using the calculated truck load test principal components. The optimized clustering model detected the outliers that are caused by heavy truck loads; clustering result is detailed. In summary, a simple data model that is able to find any known data signature such as truck load test in the daily measurement data is proposed. The proposed method is part of an ongoing effort in University of Cincinnati Infrastructure Institute to use the correlation between collected readings from different members of a bridge in order to interpret abnormal trend changes in the measurement data.
© (2013) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
M. Norouzi, V. Hunt, and A. Helmicki "Abnormal behavior detection in the Jeremiah Morrow Bridge based on the long term measurement data patterns", Proc. SPIE 8695, Health Monitoring of Structural and Biological Systems 2013, 869536 (17 April 2013);


Back to Top