Contour Methods for Deepwater Riser Systems

Abstract

This is a master thesis about environmental contour line methods for extreme responseanalysis of deep water marine riser system. The marine riser systems playsan important part of oil and gas exploration and production systems at sea. Withdierent applications and congurations, from shallow to deep water, accurate estimatesof extreme responses are signicant for the nal riser designs.In the thesis a steel caternary export have been analysed. The riser is connectedto a Tension-leg platform situated in the harsh environment Norwegian sea at 1200meters water depth. Sensitivity studies have been performed to determine importantcalculation parameters such as number elements, time steps and number ofrandom seeds.To identify the extreme sea states within a 100-year return period, the environmentalcontour line method have been used. Environmental contour lines are preferredfor non-linear structural problems as they saves computational eort compared tothe traditional full long-term analysis. Studies of both the traditional contour linemethod and a new contour line method have been undertaken. The traditionalmethod performs Rosenblatt transformations of the probability distributions ofobserved sea states to obtain the contour line. While the new method creates newsea states trough Monte Carlo simulations of the same probability distributions.Projections of the new sea states are used to identify the boundary sea states alongthe contour line.The contour lines were create from observed sea states at two areas on the Norwegiancontinental shelf, Statfjord and the Halten Bank. At both location was itfound that the traditional method gave larger signicant wave height than the newmethod, the traditional method had also the largest spectral peak periods. The seastates close to, and at, the largest signicant wave height around the contour lineswere analysed considering extreme responses in the riser. Each sea state was randomseeded 25 times. A 90% percentile of a Gumbel distribution for the responseswas used to determine extremes. The new contour line method gave almost 1%lower extreme responses in eective tension and bending moment compared to thetraditional method. Also the minimum eective tension was in the new methodsfavour.The dierence between the two contour line methods was not signicant, and uncertaintieswere introduced by irregularities along the new contour line method. Moreresearch on these irregularities and the percentile levels is suggested as importantfurther work.