Design load factor structural reliability analysis is critical for pipeline postlay OOS design to mitigate global UHB for a trenched and buried subsea pipeline configuration operating at elevated temperature and pressure. During the detailed engineering phase it is necessary to evaluate and define any measure available to be finalised for UHB mitigation such as deep trenching selection, enhanced blanket or spot rockdumping. In order account for inherent uncertainties in the design variables, a pre-emptive SRA is normally performed for the probabilistic UHB design load factors prior to pipeline installation according to the typical trench imperfection statistics and some specified survey accuracy. As per the current practice the semi-analytical universal design curve method is used in the limit state for design load factor predictions. The SRA results will be updated once the OOS survey data become available. A rockdump schedule can then be established by FEA incorporating appropriate safety or load factors to address uncertainties in the design parameters and as-built pipeline OOS survey measurement accuracy.
This paper examines the UHB model uncertainties in the load factor and backfill cover assessment with a view to improving the SRA OOS analysis. Sources of uncertainties and variability in the UHB design are discussed first. Some disparity and inconsistency arising between the SRA and FEA models for the limit state are considered. Alternative UHB models are investigated by taking Timoshenko shear stiffness and associated deformation with pipe-soil interactions into consideration. A comparison is made with the conventional universal design curve method, the improved model and FE modelling to demonstrate the findings and conclusions. Of these, the pipe-soil interaction and its representation in the SRA limit state assessment are identified as a significant factor.