RELIABILITY OF DETERIORATED MARINE STRUCTURES BASED ON MEASURED DATA

Abstract

Reliability assessment of a corroded deck of a tanker ship subjected to non-linear general corrosion wastage is performed, accounting for an initial period without corrosion due to the presence of a corrosion protection system, and a non-linear increase in wastage up to a steady state value. The reliability model is based on the analysis of corrosion depth data. Two types of uncertainties are accounted for. The first one is related to the corrosion degradation trend as a function of time, which is identified by a sequence independent data analysis. The second uncertainty is related to the variation of the corrosion degradation around its trend, which is identified as a stochastic process, and is defined based on the time series analysis. The time series determines the autocorrelation and spectral density functions of the stochastic process applying the Fast Fourier transform. The reliability estimates with respect to a corroded deck of cargo tank of a tanker ship is analysed by a time variant formulation and the effect of inspections is also incorporated employing the Bayesian updating formulation.