Assessment of necessary airgap of a semisubmersible using a Peak-Over-Threshold long term analysis
MetadataShow full item record
The airgap of a column stabilized floating structure is the distance from a point on the underside of the deck to the water level instantly and directly below it; when the sea is not calm, it varies with time and for different locations of the deck due to the wave motions. The initial airgap, the vertical distance from the underside of the deck to the mean water level, is of critical importance to make sure that wave to deck impacts are avoided; for these can result in unwanted roll or pitch motions or even damage to the topside. Safety standards require that the probability of a wave hitting the deck corresponds to 1 in 100 years for the ultimate limit state design (ULS) and 1 in 10,000 years for an accidental limit state deign (ASL). The complexity of calculating the required airgap goes far beyond estimating extreme values of the wave heights and periods because 1) the structure moves in different ways with different waves and different directions, 2) the wave field changes below the deck due to disturbances created by the structure itself 3) the surface elevation and the responses are stochastic processes. At the same time, there are several reasons to avoid ending up with an airgap that is larger than the minimum required. The geometry of a semisubmersible that was to be built offshore Norway was given along with the transfer functions that describe its motions and the functions that describe the disturbed wave elevation under the deck. These were previously computed with a finite element analysis for 16 different directions of incidence of the waves. Besides, hindcast data from the northern north-sea over the last 61 years was accessible; this describes the significant wave height (𝐻���𝑠���), peak period (𝑇���𝑝���) and direction of 3-hour sea states. Wind waves were of interest and the effect of swell was ignored due to its negligible effect. In order to determine the minimum required initial airgap for both limit state designs, statistical analyses of the responses were done for which all of the measured 3-hour sea states were examined. A peak-over-threshold (POT) approach was adopted at first, where responses are modelled for storms with 3-hour significant wave heights above the threshold. The results are compared for many cases (different thresholds and two different POT versions) and an additional analysis was made, the latter required the estimation of extreme sea states from probability contours that were generated on the 𝐻���𝑠���, 𝑇���𝑝��� scatter plot of the data. The work also presents a measure of the storm severity for different directional sectors and the estimated extreme values of the significant wave height for various analytical setups.
Master's thesis in Offshore Technology