Data-Driven Insights into Bridge Deck Condition Dynamics Using Cluster Analysis of National Bridge Inventory Data

Transportation infrastructure is critical to ensuring safety and efficiency. Among its components, bridges play a pivotal role, necessitating effective maintenance and evaluation to preserve their functionality. However, with a vast number of bridges, constrained budgets, and limited resources, state departments of transportation face significant challenges in maintaining their bridge networks. This study analyzes bridge deck deterioration in Colorado using data from the National Bridge Inventory (NBI). Temporal trends in bridge conditions are measured with dynamic time warping, and similar deterioration patterns are grouped through agglomerative clustering. These patterns are associated with structural attributes and environmental conditions identified through a random forest model to highlight key factors such as traffic metrics that influence deterioration. The study identified 14 distinct deterioration patterns using agglomerative clustering, with a random forest model achieving 92% accuracy in predicting cluster assignments. The findings enhance understanding of bridge deterioration dynamics and support targeted, cost-effective maintenance.