Numerical investigation of the scouring around pipelines

Pipelines laid on the seabed interact with currents and waves, generating a local velocity field and pressure gradients between the upstream and downstream sides of the pipe. If the pipe is laid on a sandy bottom, the interaction of pressure gradients with the bottom can drive a seepage flow, allowing the wet sand to be dragged by the shear stress. In some cases this effect can be particularly large, washing away the sand beneath the pipeline and giving rise to some scouring underneath the pipe itself. The investigation of the evolution of the pipe/seabed configuration has a large relevance for pipeline design because a free-span occurrence and persistence may cause the vibration of the pipeline and the accumulation of unacceptable fatigue damages. Therefore, for a reliable ap- proach to pipeline design, the possibility of accurately predicting both generation and evolution of scour-induced free-spans during the first years of pipe laying has a large relevance. A review of the work on the local sour below pipelines can be found in [1-3]. The majority of the studies in literature are based on experiments. This led to a fundamen- tal understanding and to the physical modeling of the phenomenon. Nonetheless numerical modelings is necessary in the design stage. The first numerical algorithms were based on the potential flow theory [4], even thought such a theory is not able to represent the vorticity generated downstream the body of interest the results were encouraging because the front de- formation was successfully predicted. Later on, full Navies-Stokes solver with turbulent models and a sediment transport equations [5-7] were introduced. The main limits of these algorithms is in the capturing of the inception of the scouring. The use of a boundary fitted mesh either leads to the use of: (1) an initial infinitesimal gap between the cylinder and the sandy bottom or (2) two simulations: one with the pipeline (i.e. cylinder) touching the bottom and, after a threshold value, to the projection of the flow field to a new topology where a gap between the body and the bottom is applied. In both cases the time of inception of the free-span depends on the simulation parameters. Here a new treatment of the boundary of the computational domain enables the capturing of the topological change in the domain, making the prediction of the start of the free-span more reliable.