A new approach to system reliability analysis of offshore structures using dominant failure modes identified by selective searching technique

Abstract

The reliability analysis of offshore structures, as conceived by current design practice, is typically carried out at the component level with an implicit assumption that the structural system will be safe as long as all its members are safe according to the corresponding limit state equations. However, it is widely accepted that a more general and rigorous approach to the risk assessment of offshore structures based on system reliability analysis is needed, in which the reliability of a structure is estimated with respect to the system-level failure caused by potential failure modes consisting of component failures with their statistical dependence considered. To this end, the merits of a new risk assessment framework, originally developed for truss and frame structures, are investigated in this paper in view of an extensive application of this method to the offshore structural systems. The main advantage of the proposed method is that the identification process of dominant failure modes is decoupled from the evaluation process of the probabilities of failure modes and the system failure event. The identification phase consists of a multi-point parallel search employing a genetic algorithm, and it is followed by the evaluation phase, which performs a multi-scale matrix-based system reliability analysis where the statistical dependence among both components and failure modes is fully considered. In order to demonstrate the applicability of the proposed method to the offshore field, the problem of a jacket platform under an extreme sea state is considered, in which the uncertainties are assumed both in the wave and hydrodynamic models and in the material properties of the structural members. The computational efficiency and accuracy of the proposed approach are successfully demonstrated through comparison with Monte Carlo simulations.