ABSTRACT
As such, a more appropriate approach is to model the extreme value distribution of the total live-load demand instead of isolating the truck load. Such an approach is developed in Messervey & Frangopol (2008) and serves as the theoretical basis for this paper. Of particular importance and focus in this work was researching how much data (number of observations) is necessary to adequately define the extreme value live load distribution and in which timeframe (daily, weekly, monthly) peak picking should occur to observe the phenomenon. To begin investigating these questions, a simulation was constructed to approximately model traffic, wind, and temperature loads on a bridge structure. From this simulation, convergence to an extreme value distribution was confirmed by peak picking maximum recordings and a discussion of the amount of error associated with short observation timeframes was initiated. However, application of the method to a real structure was lacking. This is the purpose of the study presented herein.
