Time-Variant Ultimate Reliability Analysis of Jacket Platforms Considering a New Probabilistic Corrosion Model for the Persian Gulf

Corrosion is identified as one of the most important deterioration factors for structural integrity of offshore platforms. For reliability analysis of these platforms, a probabilistic model for prediction of long-term corrosion loss as a function of time is essential. The purpose of this study is to propose a novel model for steel corrosion of jacket platforms in the Persian Gulf region. Field measurements for members in seawater are collected and statistically analyzed to identify the probability function for corrosion loss at different times. A new probabilistic model with time-dependent parameters is suggested, based on the statistical analysis results. Application of above-mentioned model in the reliability analysis of jacket platforms is investigated by introducing a new reliability analysis framework. This framework is a general solution for probabilistic analysis of jacket platforms with several stochastic variables which can be used for the platforms with different configuration and loads. In this framework, direct analysis is performed in each stage of first-order reliability method (FORM) instead of using the response surface method which is a common approach to obtaining the required response. This framework is applied to three jackets and the annual probability of failure $(Pf)$ over the platforms service life is computed. Comparison of results revealed that among the years beyond the platform design life, the amount of annual $Pf$ is increased in parabolic function. Also, studying the results is illustrated that in the case of ignoring the corrosion loss as a stochastic variable, $Pf$ is estimated 7% lower than values obtained in actual condition.