Parameterized Logistic Models for Efficient Bridge Maintenance Scheduling

Bridge maintenance is integral to bridge functionality and safety. However, for authorities with large inventories, such as the Ministry of Transportation of Ontario (MTO) in Canada which manages about 3000 bridges, management poses a challenge in light of budget and resource limitations. This study proposes an efficient yet easy-to-implement framework for life-cycle bridge management that is based on parameterized logistic models. First, a maintenance limit state is defined based on the MTO’s Bridge Condition Index (BCI). Then, the probability of exceeding this limit state is related to bridge parameters via logistic regression. The results indicate that the constructed parameterized logistic models can track the maintenance probability of a bridge network (or of an individual bridge) throughout the service life given easily accessible and limited information such as bridge age, time since last major maintenance, and location. Given an appropriate maintenance probability threshold, the proposed framework can aid bridge management authorities to predict the timing of maintenance work and increase inspection frequency when approaching these times. To estimate the appropriate maintenance probability threshold that optimizes life-cycle bridge management from safety and economic perspectives, this work conducts a life-cycle cost (LCC) analysis. Finally, a case study is presented to demonstrate the framework.