Probabilistic Time–Variant Functionality-Based Analysis of Transportation Networks Incorporating Asphalt Pavements and Bridges under Multiple Hazards

Abstract

Progressive and sudden deteriorations are the main reasons affecting the functionality of a transportation network. This paper presents a general probabilistic approach in which the ensembles of regression trees (ERT) are innovatively adopted to predict the life-cycle system reliability of asphalt pavements using Monte Carlo simulations, and pavement segments are considered with bridges in the analysis, prediction, and management of the functionality of transportation networks under both progressive and sudden deterioration due to multiple hazards. Four performance indicators of asphalt pavement subjected to multiple hazards were modeled using ERT trained with the Long-Term Pavement Performance database. The specific hazard types corresponding to each pavement performance indicator for the associated ERT model training were identified. The structural performance associated with bridge superstructures and substructures was analyzed by considering corrosion, traffic loading, and seismic hazards. The proposed approach is illustrated on an existing transportation network in Pennsylvania. The essential retrofitting timing, importance measure, and retrofitting priority associated with the individual component were investigated utilizing the calculated time-variant connectivity-based functionality and resilience associated with the network. The results demonstrate that asphalt pavements have a significant impact on the network functionality and should be considered in the postevent decision-making process of retrofitting strategies.