| dc.description.abstract | An offshore pipeline during S-laying in deep water may be subject to high pressure, bending and tensile loads. In the event of loss of tension the pipe may buckle for sufficiently high external pressures and bending moments. This may lead to rupture of the pipe wall with subsequent water ingress, termed a "wet buckle". The response of pipes subject to such load-conditions up to collapse has been well described over the past 25 or so years. There has seemingly been less focus on the post-collapse region, and no systematic studies have been found addressing the formation of a wet buckle. This thesis is aimed at modeling collapse and post-collapse of X65 steel pipelines subject to the above mentioned load conditions, with the application of ductile fracture criteria to reveal if a wet buckle could occur.
There are four different branches of ductile fracture modeling still in wide use today: empirical models, void growth models, porous plasticity models, and continuum damage mechanics. The focus of this thesis lies on two contemporary empirical models, the modified mohr coulomb model (MMC)and a model proposed by Coppola et al.
The commercial finite element software LS-DYNA is used to model collapse. Ductile fracture criteria have been incorporated in a user defined material. The return mapping algorithm is applied for the stress update for Von Mises plasticity with isotropic hardening.
Comparisons of the present finite element modeling applying the user defined material are made with analytical and experimental results of previous researchers. Under external pressure combined with bending the pressure loading acting on the bent pipe was found to give an additional moment contribution, opposing the applied bending moment. This would cause sufficiently long pipe-sections to collapse by the boundaries. For short pipe-sections, collapse would initiate by the boundaries but "migrate" towards the middle upon further bending. A compressive axial force had to be applied to ensure collapse of the middle of the pipe for all model lengths.
The main investigation is concerned with four X65 pipes, that are plausible deep-water candidates, having diameter/thickness (D/t) ratios of 10.3, 13.7, 14.4 and 19.2. The load-conditions investigated are: pure external pressure, pure bending and bending in combination with external pressure and an axial load. In the combined load-case the pressure and the axial load is incremented simultaneously to the desired level, before bending is applied by rotating the end-caps.
The equilibrium path for pure external pressure is traced by a path following method (modified Crisfield) until just prior to contact of the pipe walls. An explicit integration scheme is then used as the pressure is increased until the buckle propagation pressure is reached. For the pure bending and combined bending-pressure loading the pre-collapse region is traced by the implicit method. If dynamic collapse is found to occur, the analysis is switched to explicit.
The interaction of bending and pressure is shown to affect the shape of the collapsed cross-section, which again affects the damage accumulation. Rupture of the pipe wall is not indicated in any of the analyses, even under extreme loading, confirming the ductility of X65 steel. Damage accumulation is highly dependent on the triaxiality ratio, and this was seen to not exceed approximately 0.7, in all analyses. A realistic scenario where the pipe of D/t=14.4 collapses subject to bending at an external pressure of 22.6 MPa is found to give a damage value of 0.52 according to the MMC-criterion, where fracture is indicated by a value of 1. The damage parameter attained a maximum of 0.81 for the pipe with a D/t=19.2, subject to bending at a pressure of 32.7 MPa; it is however highly unrealistic that such a pipe is used under this external pressure. | en |
