The effect of sloshing in partially filled spherical LNG tanks on ship motions

Abstract

This report describes two methods to analyse the effect of sloshing on ship motions. The first method applies linear potential theory on both the barge and the internal tanks. For the barge the radiation, diffraction and incoming wave potentials are solved. For the internal tanks only the radiation potential is calculated, as there are obviously no incoming or diffracted waves in the tanks. The potentials are then used to solve the added mass, damping and wave forces for the coupled barge-tank system in the frequency domain. By comparing the frequency domain results with the scale model tests results a good agreement was identified for all loading conditions. On top of this the frequency domain approach is very fast and relatively easy to set up. However, linearity is assumed while in the scale model test non-linear sloshing was observed. It is expected that non-linearity will damp out the roll motion of the vessel and it is therefore expected that for increasing wave height the frequency domain solution will overestimate the response. It was also found that both roll RAO peaks are very sensitive to the wave direction. By comparing the roll RAO’s for liquid and frozen cargo in the tanks it was found that sloshing significantly decreases the height of the main roll RAO peak and also creates a lower barge natural frequency. The second method is a more complex time domain model that is based on a coupling between Volume of Fluid solver ComFLOW and ship motions solver aNySIM. In the coupled model the motions of the barge due to waves, without internal tanks, are obtained by using linear potential flow and are calculated in aNySIM. The ship motions calculated in aNySIM are used as input in ComFLOW that calculates the more complex and potentially non-linear motions of the liquid in the spherical cargo tanks. The resulting sloshing loads are again used as input in aNySIM, creating a two-way coupling between the dynamics of the ship and cargo tanks. The time domain results showed a reasonably good agreement with the scale model tests. However, the second, sloshing induced, roll RAO peak is underestimated in the coupled model. This might be due to the type of waves used in the test, uncertainty in the RAO’s and due to the properties of the coupling between ComFLOW and aNySIM. The time domain method is time consuming and complex to set up, but offers more insight in what is really happening. It was shown that non-linear motions of the barge occur due to sloshing in the tanks. It was also found that with partially filled tanks the response of the barge is irregular when exposed regular waves. The influence of the pump tower was also investigated, which showed that due to the damping created by the pump tower the overall roll response is lower and the barge natural frequencies moves to a lower frequency.