- Long-Term Bridge Performance (LTBP) Portal (Data Extraction)
- LTBP Deterioration Condition Forecast Models
January 2019 marked the release of the Long-Term Bridge Performance (LTBP) Program’s InfoBridge™ web portal, a newly developed website for dissemination and visualization of bridge data, information, and products developed by the LTBP Program. The portal’s main purpose is to leverage the analytical capability of the highway bridge research community, and fulfill the Federal Highway Administration’s (FHWA’s) responsibility to provide transparency and ready access to data collected through Federal research programs. InfoBridge also enables bridge owners with no or limited access to bridge asset management software to manage their bridge inventories through a seamless user interface that incorporates state-of-the-art querying and visualization tools. For more information visit https://infobridge.fhwa.dot.gov/Home.
Bridge condition forecast models are an essential component for implementing a data-driven bridge asset management program by optimizing the funding allocation that is associated with bridge replacement, rehabilitation, repairs, maintenance, or preservation. The LTBP Program developed three deck condition forecast models (Base Models, Survival Models, and Machine Learning Models) and implemented them in the January 2020 LTBP InfoBridge release.
The three models represent three levels of modeling complexities. Base Models are statistically deterministic and compute the time-in-condition for each bridge type from historical National Bridge Inventory (NBI) data and apply those to each bridge for deck condition forecasting. Survival models derive condition rating transition probabilities from survival-analysis curves and use Markov chain probabilistic methods to forecast future condition ratings. Machine-learning models are developed by mining the historical NBI and climate data using a deep learning approach. The three models are being further refined to improve their accuracy. Over 335,000 bridges have at least one model associated with them. For more information visit https://infobridge.fhwa.dot.gov/Home.