Long-term fatigue analysis of risers with multiple environmental random variables in time domain

Abstract

Copyright © 2015 by the International Society of Offshore and Polar Engineers (ISOPE). The deep-water offshore engineering industry is developed greatly with the increasing demand for oil and gas. Floating structural systems are usually exposed to many types of environmental loads like wave, wind, current and so on. There are two common types of limit state criteria for consideration in the dynamic analysis, which are fatigue limit state and ultimate limit state. For ultimate limit state, extreme response prediction would be assessed by either short-term or long-term method. For the fatigue analysis of riser and mooring systems, a long-term analysis in time domain is considered as the most accurate fatigue assessment approaches. However, in the long-term analysis, all possible sea states should be considered which is rather computationally time-consuming to simulate numerous sea states by classical numerical or Monte Carlo method, especially when more environmental random variables are taken into account. The conventional way for a long-term fatigue calculation recommended by design codes is the practice of lumping of sea states into a small number of manageable bins. However, there are no explicit guidelines for blocking strategy and the precision of this method cannot be guaranteed. There are some researchers studying this long-term fatigue problem, but considering only two environmental random variables: significant wave height and spectrum peak period. However, there appear to be no precedent studies for the long-term fatigue analysis of mooring and riser systems considering more than two environmental random variables. Hence, a proposed simulation approach based on State-of-the-art advanced response surface method is explored in this paper to assess the long term fatigue damage of flexible risers in deep water condition considering five environmental random variables: significant wave height Hs, spectral peak period Tp, mean wind velocity Vw, mean current velocity at the sea surface Vc and wave direction θ. The results show reasonable accuracy compared with numerical method with the using of response surface method, but with much higher computational efficiency.