ABSTRACT

ABSTRACT: The initial (i.e., at time t = 0) reliability index, the time-variant live load, and the resistance deterioration processes are some of the most important factors in conducting a reliabilitybased life-cycle analysis of a highway bridge structure. In such an analysis, at least for a newer structure, there is likely more confidence in the geometry and material properties of the structure that determine its capacity than there is in the various loading conditions and scenarios that will place demand upon the structure. Structural health monitoring (SHM) offers a potentially powerful means to obtain site-specific data. However, questions that must be addressed are: what information to collect? how often? and how it should be processed? This paper examines the potential of utilizing the statistics of extremes to answer these questions. By using on-site SHM and observing only the maximum daily peak strain values over time, it is determined that one can successfully modify an initial estimate of truck weight and volume to determine the actual distribution and volume observed at the site. Although a newly developed idea in this work and specific to an initial assumed Gumbel distribution, the method shows interesting potential in the monitoring and assessment of structural systems.