The Prelude Floating Liquefied Natural Gas (FLNG) facility is moored with an internal turret allowing it to perform offloading operations of liquefied natural and petroleum gas products. It does so in either a Free Weathervaning (FW) mode, i.e. by allowing the unit to rotate according to environmental loads, or in a Thruster-Assisted (TA) mode, i.e. by using the stern thrusters to maintain a fixed heading deemed preferable for the entire operation, or a particular phase.
An accurate estimation of the various environment effects, in terms of forces on the FLNG and LNG carrier, is critical to ensure a correct prediction of its heading or the required thruster forces, depending on the selected operating mode. The predominant loads driving the weathervaning behavior are wind and current loads.
These loads have been estimated from wind tunnel tests during the engineering phase. Since the Prelude FLNG has been installed on-site, field measurements have provided an opportunity for comparison and shown some differences with the numerical predictions based on the estimated loads, prompting a need for verification of current loads by an independent method. For the Prelude FLNG application, current loads play an important role due to facility size and significant tidal currents.
It has been shown in some previous studies that wind tunnel tests for a model of under-water geometry may underestimate current loads compared to those on a full-scale vessel. There is a boundary layer along the wind tunnel floor in wind tunnel tests, while the current profile is relatively uniform over the hull draft in the real ocean condition.
Moreover experimental tests present some additional drawbacks: they are performed at a reduced scale (1:225), the Reynolds number is lower than full-scale even with a large wind tunnel speed, and it is difficult to model the long (150m full-scale) Water Intake Risers (WIR) extending below the hull bottom.
In order to investigate these effects, state-of-the-art full-scale CFD simulations were performed for the Prelude hull and WIR. The test program included different current speeds and directions, and several sensitivity studies: Reynolds number effect between model- and full-scales, effect of current speed profile (comparing uniform and boundary layer profiles at model scale), effect of FLNG rotation in yaw, impact of unsteady current, and presence of marine growth.
Extreme dimensions of Prelude FLNG and requirements for accuracy of this study called for the CFD calculations to be performed on the High Performance Computing (HPC) clusters - Stampede2 and Frontera - at the Texas Advanced Computing Center (TACC), which are both amongst the world’s largest supercomputers.
This paper describes the assumptions and challenges of the CFD study and discusses the results of the main program and various sensitivities. The main conclusions and lessons learnt are also discussed.