Reliability-Based Modeling of Bridge Deterioration Hazards

This paper presents the results from a study on the deterioration patterns of highway bridges, using estimated sojourn times of bridge components in various deterioration states (condition). With emphasis on the age that the bridge leaves a specified threshold condition state, the uncertainties in the estimated times are best fitted as Weibull probability distribution functions. Using both complete observations' data and the right-censored data from a 13-year span condition historical data for the bridges, the maximum likelihood estimates are obtained for parameters of the fitted distributions, and reliability analyses are conducted for the various categories of bridges, by the type of roadways carried (interstate, noninterstate, and local roads), and the material type, i.e., cast-in-place concrete decks; prestressed concrete superstructures; steel superstructures, etc. The survival and hazard functions are developed and interpreted, including a combination of the deterioration hazard functions with the hazard functions of the effects of hurricane winds. It was observed that all the bridge categories deteriorated faster with age (nonexponential times), and bridge components located on the interstate roadways are deteriorating faster than similar bridges on noninterstate roadways. Asset worth of bridges are indicated, based on the reliability estimates.