National Bridge Inventory Data-Based Stochastic Modeling for Deck Condition Rating of Prestressed Concrete Bridges
Sahar Hasan, Emad Elwakil
Abstract
About 9% of bridges in the United States were classified as deficient bridges at the beginning of 2018 with about $123 billion needed for bridge rehabilitation. The bridge decks represent the highest budget associated with bridge maintenance because they deteriorate faster compared with the other components, because of direct exposure to traffic and harsh climate changes. The subjectivity in determining the condition rating is an imprecise process and may significantly affect the maintenance process, which may vary from one inspector to another. Moreover, most research works in prestressed concrete bridges condition ratings have focused predominantly on modeling and have neglected to study the individual effect of geometric variables with excluding the impact of aging and maintenance on the condition rating. The paper’s objectives and proposed contributions are investigating and modeling the impact of explanatory variables on deck condition rating apart from aging and maintenance actions. The findings highlight the design’s contribution to reducing the decline of a bridge condition rating. The stochastic regression analysis has been used to propose a realistic deck condition through a probability distribution. Four models have been developed using the National Bridge Inventory (NBI) of California, and results showed a satisfied coefficient of determination. The developed models have been validated with satisfactory results of 87% using the Average Validity Percentage Method. The developed models will help highway agencies make better decisions regarding future maintenance plans by prioritizing the bridge’s maintenance.