Abstract
As exploration for hydrocarbon resources venture into deeper waters, offshore floating structures are increasingly required to be stationed at sites of highly uncertain environmental loading conditions. Driven by high historical mooring failure rates of severe consequences, the need for effective long-term structural reliability methods arises for mooring lines. However, system nonlinearities, high problem dimensionality, and the diversity of conceivable failure causalities their extremely low probabilities complicates the analysis. Variations on the Monte Carlo approach are robust in addressing these challenges, but at the expense of high computational costs. In this study, distributions of environmental parameters and their correlations are modelled into a joint probabilistic description. By classifying conceivable sea states across the domain, an efficient uniform sampling scheme is presented as an efficient means of assessing long-term reliability against extreme events. The proposed method was performed on a floating production unit case study situated in the hurricane-prone Gulf of Mexico, exposed to irregular wave loads. The analysis was found to provide probability estimates with negligible bias when validated against subset simulation, with significant variance reduction of mean estimators by eliminating the need to over-simulate non-critical environmental conditions. The resulting sampling density has an added advantage of being non-failure specific, enabling system reliability assessments across multiple modes and locations of failure without the need for re-analysis.
