AbstractsMathematics

For the assessment of safety against free drift (loss of position) of semi-submersibles, the mooring system of a semi-submersible is introduced, and the mooring line tension is looked into based on the laboratory model test and the numerical simulations of Veslefrikk B platform. The work is mostly about the prediction of the mooring line tensions. Based on the model tests data, the characteristic mooring line tension is obtained, the breaking strength is the characteristic mooring line tension multiplied by safety factor. Then both 10^-2 and 10^-4annual extreme mooring line tensions are estimated. By fitting the maximum 10^-2 and 10^-4 annual extreme mooring line tensions to a Gumbel distribution, the probability of any extreme mooring line tension can be obtained according to the Gumbel distribution. Thus the single mooring line failure probabilities for different given safety factors can be determined. In the time domain simulation program SIMO (MARINTEK), the mooring line tensions are simulated under the same environmental conditions as the model tests, and the same analysis procedure is carried out to the numerical simulations. At last, the full long term analysis is carried out by the numerical simulations. Large numbers of sea states are simulated. For every sea state, there are 10 seeding tests, and the 3-hour extreme mooring line tensions are fitted to Gumbel. Then the response surfaces of Gumbel parameters are defined by the MATLAB 4 gird data method. The long term wave statics of the Veslefrikk field are given in the Veslefrikk Metocean Desgin Basis report [10]. After that, the long term probability for certain mooring line tensions can be calculated by integrating the conditional distribution of the 3-hour maximum mooring line tensions for the given wave characteristics times the joint distribution function of the significant wave characteristics in the critical area of Hs and Tp. The long term distribution of the extreme mooring line tensions is finally determined. And the 100-year and 10000-year extreme values are estimated.