A Bridge Structural Health Monitoring and Data Mining System

Structural health monitoring (SHM) is becoming a more widely accepted way to improve bridge management. In 2005, a fiber optic SHM system was developed and deployed by the Iowa State University Bridge Engineering Center to continuously monitor bridge performance under ambient traffic loads and to detect potential gradual deterioration or sudden damage, specifically for Iowa’s fracture critical bridges. Strain time history data collected by this system were utilized to construct a baseline model that is based upon extreme-matching distribution. Structural responses deviating from the baseline distribution are considered as indication of damage or degradation. As a means to improve the damage/deterioration prediction capabilities of the above mentioned system, it is postulated that the dispersion of the extreme-matching data could be minimized using truck (position, type, etc.) information. Thus, techniques for the determination of detailed truck information are being investigated. Finite element analysis was carried out to verify the proposed truck detection and damage detection algorithms. In this paper, the SHM system, relevant autonomous data mining results, and numerical verification are presented. Moreover, ongoing efforts in estimating the truck geometry/type, weight, and velocity are described.