Bridge Condition Assessment Using Supervised Decision Trees

Bridges are at risk from aging, fatigue and deterioration processes. Many countries are facing with large stocks of existing bridges approaching the end of the service life and the preservation of their structural performance and functional adequacy is a priority for administrations, public authorities, and decision makers dealing with bridge condition rating and infrastructure management. Visual inspections are at the base of an effective and reliable bridge condition assessment. Inspection strategies and procedures determine how results are returned, stored and managed leading to the formulation of different bridge condition indicators. The collection of information over time provides a great amount of bridge data which can be properly elaborated to get useful insights for supporting the decision-making process. Classification tools, such as Decision Trees (DTs) can be exploited to prioritize maintenance and rehabilitation interventions within the transportation network. This paper presents the application of a supervised DT for the assessment of bridge condition. The proposed approach is applied to classification in Good, Fair, or Poor status of a stock of existing bridges located in California. The DT is trained using visual inspection results stored in public United States National Bridge Inventory (NBI).