A probabilistic strategy to evaluate scour around bridge deepwater foundations considering a reliability assessment

Currently, an increasing number of deepwater foundations have been built in marine regions, and most of them have to deal with scour problems. Recent design codes of superstructures usually adopt the probability-based method, in which loads are treated as probabilistic parameters. However, for the design of substructures, the impact of scour is generally regarded as a deterministic value represented by the estimated scour depth. This inconsistency in the design strategies brings difficulties in considering the uncertainties associated with scour effects. Fundamentally, this is due to the challenges in the scour estimation using a probabilistic method. Based on an augmented scour database analyzed using the artificial neural-network (ANN), this study quantifies the uncertainties of four types of scour predictive methods. Based on this, it proposes a probabilistic approach for scour estimation according to different cases. Six typical conditions and two practical cases were analyzed using the probabilistic approach. The overall reliability index showed that all these existing scour estimation methods have certain limitations, including inconsistent reliability and unreliable scour depth predictions. To prevent these problems, a concept based on the probability of nonexceedance of a given scour depth has been proposed, which was determined by different target reliability requirements