Reliability-Based Methodology for Design and Evaluation of Concrete Bridge Decks

The bridge deck serves as a critical component of the overall bridge structure whose performance attracts great attention as it directly impacts public safety, infrastructure planning, and overall functioning of transportation networks. Bridge owners pay extensive budget on monitoring and maintaining the bridge decks to ensure their safe loading carrying capacity. While the AASHTO Load and Resistance Factored Design (LRFD) Bridge Design Specifications mandate a design life of 75 years, the actual service life of bridge decks frequently falls short due to various factors such as heavy traffic load, unsatisfactory design, environmental stressors, etc., posing particular challenges to infrastructure management and public safety.

This dissertation offers an in-depth study of load spectra, design provisions, and the deterioration of bridge decks. Nationwide weigh-in-motion data, encompassing over two hundred sites from 36 states across the United States, was utilized to scrutinize the load spectra affecting bridge decks. The analysis revealed that bridges, regardless of regional differences, are consistently subjected to overweight trucks and extreme axle loads.

Additionally. The current design specifications are critiqued for lacking reliability calibration and rigorous substantiation. This research assesses the design provisions from three angles: design load, safety level (i.e., reliability index), and multiple presence factors (MPF). It was established that design loads fail to accurately represent and accommodate the real load spectra. The reliability indices for negative moment regions align closely with the target level, while the reliability indices highlight a potential safety concern at positive moment region and hereby emphasize the need for more refined analyses to provide justifications. Regardless of the shortcomings, a positive aspect of current design provisions is the inherent conservatism of the MPF.

Finally, by integrating nationwide National Bridge Inventory (NBI) data and climate data from 49 states, this study identifies and ranks the factors contributing to bridge deck deterioration using a machine learning-based framework. Climate factors, traffic attributes, and bridge configurations are all regarded as critical determinants of the bridge deterioration rate. The study culminates in establishing the correlation between the service life of the bridge deck and the truck load spectra, revealing that both truck volume and weight can adversely affect deck service life.