Simulation of drilling riser disconnection - Recoil analysis

Abstract

The emergency disconnection system and recoil analysis is required for every offshore drilling unit. Situations can occur where the vessel needs to disconnect from the well, it can be to large forces that are being transferred to the wellhead or that the vessel is unable to maintain its position over the well. When the tensioned riser is released between the BOP and LMRP it will accelerate upwards due to released tension and unbalanced force from the tensioners, this is referred to as riser recoil.The riser tensioner system is essential for understanding the recoil and it is presented in this thesis. This system gives the force variation when the riser retracts and contains the shut-off valves used in the anti recoil system for slowing down the riser.This thesis has focused on the use of SIMA RIFLEX as the tool for making a complete recoil analysis. Modelling issues are discussed on how well the physical phenomena of the recoil can be implemented in RIFLEX. Special attention have been given to force variation and damping in the riser tensioner system. Mass loss and friction forces when the high density mud inside the riser discharges. Slowing down the riser with the anti recoil system. None of these issues can be modelled directly in RIFLEX, and requires pre processing and simplifications. Two models were developed for the use in RILFEX, one for 500 meters water depth and one for 1500 meters water depth simulating a drift-off scenario. Impact between the BOP and LMRP is an issue if the riser does not achieve enough lift off after disconnection. A worst-case scenario was set up for the 500 m model in irregular waves. No impact occurred for different disconnection timings in the selected wave. However it was found that an impact could be plausible in larger waves. In the drift-off simulation resulting bending moments on the BOP and wellhead is of focus.The built in slug model in RIFLEX was attempted used for modelling of the mass loss. It was found that the slug model does not work for a complex riser, and an alternative model was developed. By specifying dynamical nodal forces in the global system, forces can be saved to the nodes of the riser. Then both the mass and the force representing the mass loss will be saved to the same nodes, but in different matrices. The alternative model provided a good lift off from the BOP, but does not change the actual mass of the system. Compression will be another problem induced by the forces lifting the riser. SIMA RIFLEX proved to lack some modelling options to serve well for a recoil analysis.