Design Approach for Turret Moored Vessels in Highly Variable Squall Conditions

This paper focuses on examining the response of a large turret-moored FPSO or FLNG vessel in squall conditions and presents a novel and statistically robust response-based approach for the derivation of squall governed design loads. Turret mooring arrangements are typically used as a permanent mooring for FPSO and FLNG vessels, usually in deeper waters and in remote areas where storage capacity is required. The weather-vaning capability makes this type of mooring suitable for many types of environment; however, in tropical environments where metocean conditions are otherwise relatively benign, squalls can dominate aspects of design. Squalls are mesoscale convective systems that cause rapid increases in wind speed and are often associated with large changes in wind direction Squalls are highly variable in both their wind speed and direction profiles and typically uncorrelated with the wave and current conditions; hence the impact of squalls can be difficult to predict and a statistically robust industry standard design approach does not currently exist. Where squall events are the design drivers for mooring arrangements they require particular focus due to the industry’s imperfect knowledge of squall intensity and frequency coupled with the particularly high inter-annual variation of squalls at most locations. To understand the design criteria for a squall environment the horizontal motions of the turret-moored system are modelled using a simplified time-domain approach, which makes considerable simplifying assumptions. This significantly reduces computational time, allowing the analysis to be conducted efficiently for a large range of both squall conditions and associated environmental (wave and current) conditions, including by season and direction. An extreme value analysis approach is then applied to all of the maximum values of the desired response for all combinations of squall and associated conditions, by selecting a number of bootstrap resamples, each the size of the number of squall events. The bootstrap resamples are selected on the probability of occurrence of the associated conditions. A generalised Pareto model can be automatically fitted to all bootstrap resamples and for every return period a user defined number of simulations is performed to estimate the most probable or percentile estimates of the desired return period. Finally, a back calculation of the transient design conditions at the desired return period can then be made and applied into the standard design process. This ensures statistically robust design conditions for squall design loads and deriving a few statistically robust design squalls and associated conditions minimises later analysis. Importantly the approach allows the users to propagate the remaining uncertainty estimates into subsequent robustness analyses.