Dynamic performance of integral buckle arrestors for offshore pipelines. Part II: Analysis

In the second part of this study we present models for simulating the quasi-static and dynamic propagation and arrest of buckles in pipelines. The models are developed within the framework of the nonlinear finite element ABAQUS and are used to simulate the quasi-static and dynamic arrest experiments in Part I. They are based on finite deformation kinematics and properly treat the contact that develops in the collapsed tube behind the propagating buckle. In the quasi-static model, the tube and arrestor materials are modeled as J2 flow theory solids with isotropic hardening. In the dynamic model, the rate dependence of SS-304 is assumed to exhibit an overstress power–law dependence. The pressurizing medium is assumed to be vacuum. The dynamic model is shown to reproduce accurately the conditions of steady-state buckle propagation as well as the dynamic engagement of a buckle with an arrestor. For buckle velocities of interest, the buckle profile was found to have sharpened considerably compared to the quasi-static one. The numerical results showed the same dynamic enhancement of arrestor performance observed in the experiments. When the much sharper dynamic buckle profile engages an arrestor, it initially induces to it and to the downstream tube significant reverse ovality. This tends to obstruct and delay the flattening of the arrestor and tube which is the cause of the crossing of arrestors with quasi-static efficiencies of less than 0.7. The net result is that a higher pressure is required to cross the arrestors when the buckle is running. As in the experiments, the biggest dynamic enhancement was found to occur for arrestors of efficiencies in the range of 0.35–0.6. Based on the results of this study it is concluded that quasi-static design procedures for integral buckle arrestors proposed previously are conservative.