Modeling the Risk of Advanced Deterioration in Bridge Management Systems

The Florida Department of Transportation (DOT) has developed a set of risk models for its bridge management system; the models are built into its existing Excel-based project-level and network-level decision support tools and are intended for eventual use in AASHTO's Pontis 5.2. One of these new models is an analysis of the risk associated with advanced deterioration. This model extends the reach of Florida's existing deterioration models to estimate the likelihood of service disruption if a badly deteriorated element should be unrepaired. The Florida DOT maintains data about demolished and replaced bridges as inactive records in its Pontis database. The condition and characteristics of the removed bridges were statistically analyzed to explore the reasons for the end of each bridge's life. On the basis of the available data, including information on load posting and reconstruction, the likelihood of service disruption was reasonably quantified. The best models developed in the research used a combination of linear and lognormal forms and reflected the buildup of risk caused by repeated Markovian deterioration as well as the typical delay that occurred between the observation of the deteriorated conditions and the resulting action to replace or reconstruct the bridge. As a by-product of the research, a linear failure probability model was developed; the model is suitable for Pontis releases up to 4.5. This model will be helpful to bridge owners who do not have their own source of estimates for the probability of bridge element failure.