Prediction of Extreme Tether Tension for a TLP by the AUR and ACER Methods

Abstract

The paper presents a study of the extreme value statistics related to measurements on a scale model of a large tension leg platform (TLP) subjected to random waves in a wave basin. Extensive model tests were carried out in three irregular sea states. Time series of the motion responses and tether tension were recorded for a total of 18 three hour tests (full scale). In this paper we discuss the statistics of the measured tether tension. The focus is on a comparison of two alternative methods for the prediction of extreme tether tension from finite time series records. One method is based on expressing the extreme value distribution in terms of the average upcrossing rate (AUR). The other is a novel method that can account for statistical dependence in the recorded time series by utilizing a cascade of conditioning approximations obtained by defining the average conditional exceedance rates (ACER). Both methods rely on introducing a specific parametric form for the tail part of the extreme value distribution. This is combined with an optimization procedure to determine the parameters involved, which allows prediction of various extreme response levels.