Probabilistic Assessment of Hull Girder Strength and Stiffness

The paper summarises recent work on a probabilistic prediction of hull girder strength and stiffness. This addresses a requirement within the UK Ministry of Defence to introduce structural reliability based methods for naval ship structures, which can have benefits in design as well as inservice assessment. A Monte Carlo simulation scheme, which samples a number of probability distributions, has been developed for application with an ultimate strength prediction tool. This tool uses 2D representation of hull girder sections. The section models are divided into regions that are assumed to have the same statistical properties within each simulation cycle. It is demonstrated that a lognormal distribution describing the ultimate strength of a hull girder section can be derived with relatively little computational effort. The results indicate that whilst standard deterministic strength predictions fall below mean values, there is a significant probability that actual hull girder strengths may be lower. This is of the order of 5%. Modelling uncertainties associated with the simplified strength prediction method are also outlined and quantified. The same method can be used for calculating hull section modulus. Results can be used directly when transforming measured strains to applied bending moments as part of work on load prediction and validation of hydrodynamic tools. The work presented forms part of a larger research programme that aims to produce recommendations for a limit state based assessment of hull girder strength against wave induced bending.