Assessment of Uncertainties in Estimated Wellhead Fatigue

During subsea drilling, the motions of the Mobile Offshore Drilling Unit (MODU) and riser cause cyclic bending moments on the wellhead. This may lead to fatigue failure of the component, which can have severe consequences, and there are currently no international standards on how to carry out a wellhead fatigue damage assessment. Through the Joint Industry Project (JIP) Structural Well Integrity, a general wellhead fatigue analysis method has recently been proposed, where the fatigue damage is estimated from a global and a local finite element analysis. There are however several uncertainties regarding the modelling of the systems, which can have significant effects on the estimated fatigue life. In this thesis, some of these uncertainties are identified, and their effect on wellhead fatigue is studied.A short description of a typical subsea drilling system is given, and relevant theory on the topic is presented. The theory consists of an introduction to dynamic response analyses of risers and fatigue life calculations for marine structures, as well as a summary of the general wellhead fatigue analysis method described in the current method.A global riser model is established in Sima/Riflex, which is a software developed for analysis of slender marine structures. A local wellhead system model is also established in the finite element program Abaqus, and the fatigue damage in a typical North Sea well is assessed for a one year long historical operation. The first uncertainty to be studied is the modelling of interaction between conductor and soil. This is usually done using p-y curves, which relate the lateral pressure from the soil to the displacement of the structure. The method describes how to construct p-y curves for both static and cyclic loads, where the static curves are commonly used. In this thesis, the analyses are also conducted with cyclic soil springs. Another modelling aspect with respect to soil interaction investigated in the thesis is soil damping, which is not included in the standard method.The standard is to calculate drag force on the riser in the global analysis using the maximum projected diameter of the main tube and auxiliary lines for all wave headings, which results in maximum drag. The wave heading will vary during an operation, and the effect of using the projected diameter for a flow perpendicular to the one giving maximum drag is therefore evaluated. The fluid particle velocities and accelerations used to calculate the forces on the riser are usually found from the undisturbed incoming wave, while they in reality will be affected by disturbance from the MODU. To study this effect, diffraction from a simplified MODU geometry is calculated, and included in the global analysis. In addition, directional data on weather and MODU response is used to get an indication of how significant conservative directional assumptions are for the estimated fatigue damage, as long created waves in the most unfavourable direction usually are applied.The results from the fatigue assessment conclude that the conductor deflections for the studied well are too small for cyclic p-y curves to have an effect on estimated fatigue life. A less stiff wellhead system could get a reduction in fatigue damage, but displacements of that order may not be realistic. The results also show that reduced drag and diffraction from the MODU have relatively little effect on fatigue damage. A reduction in drag force of 23.5 % gives an increase in total fatigue damage of 9 - 10 %. Diffracted wave kinematics result in a reduction in estimated fatigue damage of 8 % for head sea waves, while a 6 % increase in fatigue damage is observed for beam sea waves. Material soil damping is seen to have a significant effect on wellhead fatigue, as the total estimated damage is reduced with 34 - 40 %, with a reduction up to 90 % for certain sea states. There are however great uncertainties associated with the calculated damping values. The results also show that using directional data can give large reductions in estimated wellhead fatigue. The estimated fatigue life of the wellhead is found to be over three times longer for head sea compared to beam sea, and a simplified spreading of the fatigue damage around the circumference of the hotspots is seen to reduce the estimated damage by 25 % in the applied environmental conditions. It can,however, be difficult to justify non-conservative directional assumptions for future operations