Content area
The New Mexico State University (NMSU) Bridge Inspection Program (BIP) supports the New Mexico Department of Transportation (NMDOT) by performing the routine inspections for 25% of the bridges in the state. The principal challenge NMSU BIP faces is the management of the inspection data and inspection route planning, due to the manual handling of the data and manual scheduling of inspections.
This thesis proposes the development of a pilot Bridge Management System (BMS) for the NMSU BIP that improves the inspection planning and data management using data-driven strategies. The BMS was composed by integrating Information Management, Data Analysis and Decision Support. The Information Management developed the BIP dataset containing the National Bridge Inventory (NBI) information of the bridges assigned to BIP.
The Data Analysis applied two approaches: Machine Learning classification models to provide the visualization of the bridge conditions along the state using geographical and NBI information associated to the condition states of the bridge; and metaheuristic optimization algorithms to determine the optimal inspection route.
The Decision Support integrated the priorities the BIP selected for optimizing the inspection routes into a utility function, that considered the ratio between the total distance and total cost related to the optimal route, to evaluate the algorithms performance. The evaluation showed that the Bee Colony Optimizer Algorithm provided the most balanced performance based on the utility function and computational performance. The findings of this research provide a practical framework for improving bridge inspection planning at NMSU BIP.