Development of Adjustment Factors and Load Ratings via Statistical Analyses of the National Bridge Inventory Database

Details

• Creators:
  McVey, Isaac ; Husic, Admin ; Collins, William ; Lequesne, Rémy ; Roundy, Joshua
• Corporate Creators:
  University of Kansas. Dept. of Civil, Environmental and Architectural Engineering
• Corporate Contributors:
  Kansas Department of Transportation. Bureau of Research ; United States. Department of Transportation. Federal Highway Administration
• Subject/TRT Terms:
  Adjustment Factors Bridge Performance Bridges Inventory Control Load Factor Serviceability Statistical Analysis Structural Health Monitoring
• Publication/ Report Number:
  K-TRAN: KU-22-2
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States ; Kansas
• Edition:
  Final Report: July 2021 – February 2024
• Corporate Publisher:
  Kansas Department of Transportation. Bureau of Research
• Abstract:
  This study developed a new approach to establish baseline load ratings for bridges in Kansas without plans using data from the National Bridge Inventory (NBI). The approach is comprised of linear regression models to estimate load ratings for bridges with a condition rating of 8 or higher and adjustment factors to lower the estimated load rating to account for bridge condition ratings of 7 or lower. This approach beneficially establishes baseline load rating estimates for structures without prior ratings and secondary load ratings for bridges with prior load ratings to identify outliers and potential errors. The adjustment factors can be used to adjust load ratings obtained by any method to account for bridge condition if the condition was not specifically integrated into the analyses. Both the linear regression models and condition adjustment factors are designed to reflect trends among Kansas bridges within the NBI, not engineering judgment. This approach answers the following question for a given bridge: Knowing nothing more about the structure than what is available within the NBI, what is the expected rating based on similar bridges in similar condition within Kansas? The proposed linear regression models include bridge age, modeled design load, structure kind (construction material), structure type (truss, girder, etc.) and deck width because, among variables reported in the NBI, these were most closely correlated with load rating. The adjustment factors were developed based on the median reported load rating for bridges with various condition ratings, and uncertainty was estimated using a bootstrapping simulation. The proposed models demonstrated satisfactory performance, capturing approximately half the variance observed in the data for the Inventory (R2 = 0.50) and Operating (R2 = 0.49) Ratings. Further validation and refinement, inclusion of additional predictors, and exploration of alternative methods are suggested to improve accuracy and applicability.
• Format:
  PDF
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha-512:05f9ac609365bddd1479ef80e4ae954bd7683052df17e3fcecec4de49a0b15101f526e70e2cf2c1bcc970d2395ded54a209a7a5cf1646e9eb8bf61eeb9aa0547
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/88863/dot_88863_DS1.pdf
• File Type:
  [PDF - 2.02 MB ]