Analysis of an Offshore Jacket subjected to Supply Vessel Impacts

Abstract

In this master s thesis an offshore jacket platform subjected to supply vessel impacts is analysed. Since the supply vessels have increased in size and have been reinforced through modern ship design it has been of significance to study ship impacts with larger ships.

The collision scenarios are chosen based upon damage potential to critical members such as risers and conductors and the damage potential to structural integrity. Two stern collisions against risers, a side collision against the platform leg and a bow impact against a conductor area have been analysed. The supply vessel used in this master s thesis had a displacement of 7500 tonnes and according to new ALS-requirements the speed at impact should be 3m/s for bow impacts and 2m/s for side and stern impacts. Hence, in view of new collision requirements it is desirable to study whether the jacket platform is capable to withstand a bow-, side and stern impact of 37.12MJ, 21MJ and 16.5MJ, respectively.

Local analyses have been performed with NLFEA in LS-DYNA while the global analyses have been carried out in USFOS. All LS-DYNA analyses are decoupled, which means that rigid body motions of the ship (e.g.: change in speed, direction due to impact) are not considered. Furthermore, the LS-DYNA analyses were performed quasi-statically which means that the ship was pushed towards the jacket platform at constant speed until the internal energy (strain energy) reached the collision energy level.

Structural sub-models of the jacket platform were modelled and meshed in SESAM GeniE. The structural ship models used in the LS-DYNA analyses are the same which are included in DNVGL-RPC208. The bow-, side and stern model are all designed to be representative for an OSV with a displacement in the range of 6500-10000 tonnes. The USFOS-model was provided, and only minor modifications were done. During the LSDYNA analyses the energy absorption in both ship and jacket was analysed. In USFOS, the ship was presented as a nonlinear spring based upon the ship deformation behaviour observed in LS-DYNA. The stern collision showed that there was a severe damage potential to the risers. Due to inaccurate modelling of the riser clamps and riser flanges it was not possible to judge the risk of rupture. Inaccurate modelling of one of the clamps between one of the braces and one of the risers caused rupture of the brace, a result which is questionable. Internal pressure and temperature in the risers were not considered in the LS-DYNA analyses. Another uncertainty was that a stern corner was used, and it is therefore questionable if the boundary conditions along the geometrical symmetry plane are accurate. In the side collision analysis deformation of both ship side and platform leg were achieved. Out of a collision energy of 21MJ, 15.5MJ and 5.5MJ were dissipated by jacket and ship, respectively. The bow impact against the conductor area showed that the conductors were strong enough to crush the forecastle and deform the bow. Internal pressure in the conductors was not implemented due to the design of the conductor. Of 37MJ the ship absorbed 30MJ while approximately 7MJ were dissipated by a diagonal which deformed in a three-hinge mechanism.

Due to time limitations a mesh convergence study was not carried out on neither of the sub-models. Furthermore, strain rates were not included. The rupture and tensile fracture criteria are also mesh dependent in NLFEA and must be chosen according to calibration procedures in DNVGL-RPC208. These are the main uncertainties in the LS-DYNA analyses. In USFOS, the commands BIMPACT, MULT_IMP and SURFIMP were used. Despite challenges of capture the dissipating energy, the energy dissipation results were close to the results obtained with LS-DYNA. The stern collisions in USFOS gave reasonable results regarding fracture of braces. Risers and clamps were also affected by the deformation of the braces. In the side impact against the platform leg there were good coincidence with the results obtained with LS-DYNA. Since the conductors are non-structural in USFOS, the bow impact turned into a capacity check of the diagonal. The diagonal absorbed 9.5MJ before fracture, while the diagonal absorbed 7MJ in LS-DYNA. For all collision scenarios the damaged jacket survived the residual strength check without structural collapse. In the residual strength check, the jacket platform was subjected to a 5th order Stoke wave with a return period of 100 years.

The master s thesis has concluded that the stern corner of the ship managed to do severe damage to the risers before it hit the platform, while the conductors seemed strong enough to crush the ship. Recommendations for the SURFIMP-command are given in Appendix. Based upon observations in LS-DYNA it was concluded that it is too optimistic to set the dent width equal to the height of the contact area between the ship and the platform. The dent widths in LS-DYNA were observed to be in the range of 0.1m-0.3m, instead of 2.5m-3m which was based upon ship geometry from earlier analyses. Smaller dent widths cause a more concentrated collision which reduces the capacity of the platform leg. For ship impacts against platform legs which do not stand perpendicular to the sea surface it is recommended to choose a dent width based upon NLFEA-results rather than ship geometry.