EXTREME RESPONSE ESTIMATE OF STEEL CATENARY RISERS USING L- MOMENTS

摘要

The steel catenary risers (SCR) respond non-linearly to FPU motions, especially in extreme environments such as 100-year and 1000-year hurricanes. One of the design challenges is an accurate estimate of SCR response based on limited realizations. Due to the nonlinear nature of SCR response and the lack of enough sample data, traditional methods did not produce optimum extrema estimation. L-moments have been applied successfully in the statistical analysis of data, because they are linear combinations of probability weighted moments inherently and are less sensitive to outlying data values. Therefore, L-moments have been introduced in this paper to determine the candidate distribution of SCR responses and estimate extrema. Three typical SCRs connected to semi-submersible FPU have been analyzed using this methodology. Simulations confirmed L-moments can accurately determine the SCR responses distribution. Moreover, the extreme values predicted in this way were consistent and reliable as compared with current industry practice. Those observations have demonstrated that L-moments are robust and efficient methods in the design of steel catenary risers.