Discovering Reasons and Patterns for Bridges with Fast Deteriorating Deck Conditions

• Activity Number: 168 - Risk analysis and related topics
• Type: Contributed
• Date/Time: Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Transportation Statistics Interest Group
• Author(s): xiaoqiang kong* and zihao Li and Yunlong Zhang
• Companies: Texas A&M University and Texas A&M University and Texas A&M University
• Keywords: deck condition; interpretable machine learning; pattern recognition; deteriorating

Abstract

The deck condition of a bridge is one of the most important factors which impacts the connectivity and efficiency of the whole transportation network. The National Bridge Inventory (NBI) contains a massive amount of bridges across the United States highway system. Many relatively young bridges (less than 20 years old) across the country have poor or fair deck conditions. Meanwhile, many relatively old bridges (30-40 years old) still maintain a good deck condition without re-construction. This study explores the reasons behind these bridges with fast deteriorating deck conditions. An interpretable machine learning framework has been developed and applied to the NBI database. The results show the climate region, structure width, number of spans, deck structure, etc. are strongly associated with the fast deteriorating deck conditions. The more in-depth analysis finds that the cast-in-place concrete deck structure has the best performance. The roads with higher truck traffic generally have better deck conditions, which may suggest the road with higher truck traffic receives more funding and maintenance. Findings could help practitioners develop more effective countermeasures.