Evaluation of Steel Bridge Girders Damaged by Over-Height Vehicle Strikes Using Terrestrial Laser Scanning and Machine Learning

Details

• Creators:
  Ibrahim, Ahmed E. ; Schissler, Limo R. ; Fahnestock, Larry A. ; LaFave, James M. ; Elbanna, Ahmed E.
• Corporate Creators:
  Illinois Center for Transportation
• Corporate Contributors:
  Illinois Department of Transportation. Bureau of Research ; United States. Department of Transportation. Federal Highway Administration
• Subject/TRT Terms:
  Bridges Evaluation Girder Bridges Girders Impact Impact Damage Low-Clearance Bridges Machine Learning Point Clouds Scanners Steel Bridges Structural Damage Terrestrial Laser Scanning

  More +
• Publication/ Report Number:
  FHWA-ICT-26-009 ; ICT-26-011 ; UILU-2026-2011
• DOI:
  https://doi.org/10.36501/0197-9191/26-011
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  Final Report: 10/1/20-7/31/26
• Corporate Publisher:
  Illinois Center for Transportation
• Abstract:
  Bridge strikes caused by over-height vehicles frequently damage steel girders and can reduce their load-carrying capacity, requiring rapid and reliable evaluation to ensure structural safety. Current evaluation practices rely on manual measurements that often require traffic disruptions and simplified assessment approaches that do not fully capture the effects of damage. In addition, finite element simulations are time-consuming and computationally intensive to develop for practical rapid evaluation. This study developed an integrated, data-driven framework for the inspection and assessment of damaged steel bridge girders. A semiautomated procedure was developed to process laser-scan point cloud data and extract displacement profiles along the girder span. Machine learning models were trained and validated using a dataset generated from finite element simulations informed by field measurements of actual bridge strike incidents to predict the residual capacity of damaged girders. Explainable artificial intelligence techniques were incorporated to interpret model predictions and identify key parameters influencing structural performance. Results demonstrated the framework's capabilities in effectively capturing girder displacements, and for accurately and efficiently estimating the residual capacity, to support post-strike bridge evaluation. Measured girder deformations obtained from laser scanning can be used as inputs to the machine learning models for quantifying a girder's capacity. The predicted capacity can then be incorporated into bridge rating software to compute rating factors, enabling rapid decisions on traffic restrictions and load posting. Additionally, the interpretability results can guide engineers in identifying critical design parameters and prioritizing repair actions.
• Format:
  PDF
• Funding:
  R27-225
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha-512:e76b52f79ea4f7c76d4eb2df90ed0a0e35cb9855516821c95b74c83676626f8c6e1cc63b993413fff1ae42cd0c75ca43e09ae005828cf3030f0ad183726c222e
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/92770/dot_92770_DS1.pdf
• File Type:
  [PDF - 15.85 MB ]