Uncertainty of Long Term Fatigue Load of Subsea Well Heads

Well heads (WH) are exposed to dynamic loading from waves and floater motions through the marine drilling riser and the drill string. Risers are known to have a pronounced nonlinear response characteristic, this influences the dynamic loading of the well heads. Non-linear time-domain finite element analyses are in general required to give an adequate description of the non-linearities involved. Analyses of a large number of short-term environmental conditions considering stochastic wave loading are required to give a representative description of the long-term fatigue loading on the structure. The objective of this paper is to give an assessment of the sufficient simulation time of each short term seastate in order to get a required level of the statistical uncertainty of the long term fatigue load of the well head and conductor. The analyst can actually influence this statistical uncertainty by selecting proper analysis methodology. A statistical uncertainty measure can be used to evaluate the robustness in the estimated fatigue life. Assessment of statistical uncertainty in fatigue damage estimate is demonstrated by a case study, using typical North Sea conditions, and for varying boundary conditions. Short term fatigue damage is established by means of rainflow cycle (RFC) counting in each stationary short-term condition. It has been experienced that significant statistical uncertainties can be present in the short-term fatigue damage estimates. This is because the accumulated fatigue damage in a stationary condition normally has significant contribution from the largest stress cycles in the realization. Selection of sufficient simulation length is hence essential to obtain reliable fatigue life estimates. Applicable codes and standards for risers provide Design Fatigue Factors (DFF) to secure adequate safety against failure due to wave induced fatigue. The total uncertainty in the calculated fatigue damage comes from various sources and the DFFs in e.g. DNV-OS-F201 “Dynamic Risers” and API-RP-2RD corresponds to a certain uncertainty level in the fatigue damage estimate. A recommended target value for the statistical uncertainty of the fatigue damage estimates is given with basis in these design codes. There are no similar target values for statistical uncertainty of long term fatigue damage given in the applicable design codes for well heads and drilling equipment, hence the values from the riser codes have been selected as a starting point. The recommendations and methodologies presented in this paper, will be included in the upcoming DNV RP “Fatigue design and analysis of drilling and well equipment”, at present this is available only as a draft, ref. [1].