Supporting Bridge Management with Advanced Analysis and Machine Learning

This project seeks to provide easy access to an approximation of advanced structural analysis, allowing practicing engineers to more accurately calculate bridge load ratings without needing to perform rigorous analyses themselves. Advanced analysis can provide insight into whether or not bridge management intervention is necessary. Rigorous modeling can sometimes reveal unacknowledged capacity overlooked in traditional simplified models. This existing but unacknowledged capacity can potentially be sufficient to justify removal of load posting and deferral of bridge maintenance or replacement. Effectively managing the bridge inventory serves several strategic goals identified by the Nebraska Department of Transportation, particularly by balancing safety and fiscal responsibility. The primary objective of this research is to investigate and demonstrate the validity and usefulness of artificial neural networks (ANNs) as a supplementary tool for bridge load rating and bridge management decision-making, substantiated through validation with diagnostic bridge tests. This research will calibrate and/or refine an existing, preliminary ANN model to better serve the needs of Nebraska Department of Transportation (NDOT), by expanding the ANN training data with Nebraska bridges and integrating reliability into the ANN predictions consistently with American Association of State Highway and Transportation Officials (AASHTO) Load and Resistance Factor Design (LRFD)/R.