Ship structures are inspected regularly to reduce uncertainty associated with deterioration and help identifying optimal maintenance actions. The effect of inspections can be quantified by updating probabilistic models of the ship structure and deterioration processes with the information obtained during inspections. Most deterioration processes in ship structures, such as corrosion, are spatially distributed. In this contribution, we investigate how the spatial distribution and dependence of corrosion can be adequately addressed in Bayesian updating and how it affects the ship reliability. To model the spatial variability of corrosion in a ship structure, we apply a hierarchical spatial model. Inspections of the structure by means of thickness measurements are considered. Bayesian updating of the spatial corrosion model and the ship reliability is performed by means of the BUS approach. Finally, comparison is made with reliability estimates obtained with a model that neglect spatial variability and with the classical approach that defines separate random variables for the corrosion groups.     «

Ship structures are inspected regularly to reduce uncertainty associated with deterioration and help identifying optimal maintenance actions. The effect of inspections can be quantified by updating probabilistic models of the ship structure and deterioration processes with the information obtained during inspections. Most deterioration processes in ship structures, such as corrosion, are spatially distributed. In this contribution, we investigate how the spatial distribution and dependence o...     »