Reliability Analysis of a Bridge Deck Utilizing Generalized Gamma Distribution

Models of bridge deck deterioration with improved predictive power can provide management strategies that will optimize allocation the available, typically limited, budget transportation agencies in a more efficient manner. In turn, this lead to improvement overall condition bridges. To end, study developed novel statistical hazard model using generalized gamma accelerated failure time attributes as covariates. Bayesian inference was used estimate parameters and update these new inspection data become available. The Markov chain Monte Carlo sampling method posterior distribution utilizing both uncensored censored data. proposed approach applied approximately 30 years in-service performance inspected from 1985 2015 for than 22,000 bridges state Pennsylvania. results showed based on had high accuracy. Further, reliability different attribute values physical makeup main span structure, interaction type, rebar coating are quantified results. help improve predictions future conditions decision-making tools infrastructure management.