Reliability assessment of ship structures using Bayesian updating

This paper presents an approach for reducing the uncertainty in the performance assessment of ship structures by updating the wave-induced load effects with the data acquired from structural heath monitoring (SHM). The initial information on the wave-induced load effects is calculated based on strip theory. Bayesian updating method is used to incorporate the processed SHM data to update the wave-induced vertical bending moment modeled by the Rayleigh distribution and its extreme values modeled by the Type I extreme value (largest) distribution. Three general cases associated with updating (a) only one parameter, (b) two parameters separately, and (c) two correlated parameters simultaneously in the Type I extreme value distribution are investigated. Aging effect due to corrosion is considered in the resistance modeling. Time-variant reliabilities before and after updating are evaluated. The proposed approach is illustrated with the Joint High Speed Sealift.