New Measure to Understand and Compare Bridge Conditions Based on Inspections Time-Series Data

• Published in: Journal of infrastructure systems 2021-12, Vol.27 (4)
• Main Authors: Kale, Akshay, Ricks, Brian, Gandhi, Robin
• Format: Article
• Language: English
• Subjects: Bridge inspection; Bridges; Datasets; Environmental factors; Infrastructure; Inspection; Ratings; Reproducibility; Technical Papers; Time series
• ISSN: 1076-0342; 1943-555X
• DOI: 10.1061/(ASCE)IS.1943-555X.0000633

AbstractThe C+ score for US bridges on the 2017 infrastructure report card underscores the need for improved data-driven methods to understand bridge performance. There is a lot of interest and prior work in using inspection records to determine bridge health scores. However, aggregating, cleaning, and analyzing bridge inspection records from all states and all past years is a challenging task, limiting the access and reproducibility of findings. This research introduces a new score computed using inspection records from the National Bridge Inventory (NBI) data set. Differences between the time series of condition ratings for a bridge and a time series of average national condition ratings by age are used to develop a health score for that bridge. This baseline difference score complements NBI condition ratings in further understanding a bridge’s performance over time. Moreover, the role of bridge attributes and environmental factors can be analyzed using the score. Such analysis shows that bridge material type has the highest association with the baseline difference score, followed by snowfall and maintenance. This research also makes a methodological contribution by outlining a data-driven approach to repeatable and scalable analysis of the NBI data set.