Computational Fluid Dynamics Analysis as Applied to the Prediction of Dynamic Hookload Variation in Deepwater Drilling Risers

Abstract

This paper discusses computational fluid dynamic (CFD) analysis that has been applied to a blowout preventor (BOP) stack undergoing vertical oscillation during deployment from a floating drilling rig. CFD analysis has been used to determine added mass and to show that the drag on an oscillating BOP stack can be much more favorable than the calculations based on steady flow would suggest. The mechanism responsible for this improved drag and its importance to offshore drilling operations are discussed. Exploration drilling in deeper water coupled with the need for larger, heavier BOP stacks is causing increased utilization of hook load capacity on many drilling rigs, including the current generation of newbuilds. Thus, accurate prediction of dynamic hookload fluctuation is of very significant practical importance and it cannot be done conservatively without damping assumptions that have solid justification. For deployments to 10,000-12,000 feet, the resonant period of most drilling risers (characterized by vertical motion of the BOP stack) is 5-8 seconds. In otherwise benign sea states (low wave heights with short wave periods), even small amounts of vessel heave can induce resonant dynamic variation in hook load. Predictions of this resonant response are very sensitive to assumptions about damping. The use of drag coefficients based on steady flow can significantly underestimate the amount of damping that occurs as the BOP stack oscillates vertically, especially for small-amplitude oscillations. Previous work has shown that this effect occurs in other offshore structures as well. This convergence of resonant period and wave period is more pronounced for drilling risers run to the current water depth limit of offshore exploration (10000-12000 feet) than in shallower water, where most of the industry's experience has been gained. For deep deployments, this type of analysis can be used to substantiate more favorable assumptions about drag, when more conservative assumptions may show that riser deployment to deep wells is not practical and when arbitrary assumptions about damping may lead to very significant errors, conservative or otherwise, in the predicted dynamic load. This insight can also be applied to other large payloads deployed on long strings. Introduction For deployments to sites in water depths of 10,000 to 12,000 feet, the prediction of dynamic variation in hook load during running of a drilling riser and BOP stack requires accurate calculation of drag and added mass. Although the drag and inertial forces that act on the BOP stack are relatively small in comparison to the dynamic hook load at the top of the riser, accurate assumptions about them are important. Added mass influences the resonant period of the system (where increasing mass leads to longer resonant periods which tend to increase the dynamic response). Drag is an important source of damping. Since damping reduces resonant response, analysis that reveals larger drag forces can be used to demonstrate that deployment of a BOP stack to extreme depths is more feasible than typical drag assumptions would indicate. This work shows that the use of drag coefficients based on steady flow significantly underestimates the amount of damping that occurs as the BOP stack or riser oscillates vertically, especially for small-amplitude oscillations. This finding is of significant practical application for predicting dynamic response of drilling risers to deep depths.

Validation of Computational Fluid-Structure Interaction Analysis Methods to Determine Hydrodynamic Coefficients of a BOP Stack

“…With the advent of large sixth-generation BOP stacks with the possibility of additional capping stacks, such approximations are no longer acceptable. Therefore, detailed and state-of-theart computational fluid dynamics based analytical approaches are gaining a foothold in determining these coefficients [1,2]. Concurrent model tests can be used in blind comparisons to benchmark the calculations.…”