i-BM: An Intelligent Bridge Management Tool for Bridge and Culvert Deterioration Forecasting and Anomaly Detection Based on Physics-Guided Deep Learning

Details

• Creators:
  Banaei-Kashani, Farnoush ; Rens, Kevin
• Corporate Creators:
  University of Colorado at Denver
• Corporate Contributors:
  Colorado. Department of Transportation. Applied Research & Innovations Branch ; United States. Department of Transportation. Federal Highway Administration
• Subject/TRT Terms:
  Artificial Intelligence Bridge/Culvert Anomaly Detection Bridge Decks Bridge Management Systems Bridges And Culverts Data-Driven Data Collection Deep Learning Deterioration Failure Forecasting Freeze Thaw Durability Loads Mathematical Models Performance Physics Physics-guided Deep Learning Model Tools

  More +
• Publication/ Report Number:
  CDOT-2025-02
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States ; Colorado
• Edition:
  Final Report
• Corporate Publisher:
  Colorado. Dept. of Transportation. Applied Research and Innovation Branch
• Abstract:
  Bridges and culverts deteriorate with time and use. The deterioration process is affected by several factors, such as structural materials, structural design and behavior, daily traffic, freeze and thaw cycles, climate, pollution, temperature variation. After a certain period of time has elapsed, the deterioration processes accelerate and in a relatively short time interval the components can lose the capacity to carry the loads they were designed to support. While in the past, various data-driven deterioration models, including Bayesian models, probit model, and Markov chains are proposed in the literature to model bridge deterioration, these models either suffer from low accuracy or are too complex to be applicable. In the past we have developed AI-based deep learning (DL) models for enhanced bridge management. In particular, we focused on developing DL-based models for bridge subtyping (or bridge family generation) and bridge (and culvert) deterioration forecasting. Our results show that DL-based models for bridge subtyping and bridge/culvert deterioration outperform other existing models and can be used to effectively enhance bridge management. With this problem definition, we develop an intelligent bridge management tool (dubbed, i-BM, short for intelligent Bridge Management) for bridge and culvert deterioration forecasting and anomaly detection to be used by CDOT bridge engineers as a tool for effective bridge management. i-BM builds on and significantly extends our aforementioned prior work on DL models for bridge management in three ways: 1) integrating the DL models into a user-friendly software tool with graphical user interface and enhanced operational features, 2) developing enhanced physics-guided DL models that integrate traditional physics based bridge deterioration forecasting models with data-driven DL models for further improved performance in prediction of deterioration, and 3) introducing bridge performance anomaly detection as a new capability that allows for accurate prediction of bridge performance anomalies such as those that can lead to bridge failures/accidents.
• Format:
  PDF
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha-512:353f4f413d08c9da9951f407976ce2790c5d7210ced67416d238d5bf29d415e2c4b2119c693090dc813555cf0bc1269cbea9f0700aa07e34dfa3b860323fe992
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/89056/dot_89056_DS1.pdf
• File Type:
  [PDF - 5.54 MB ]