A comparison between Smoothed-Particle Hydrodynamics and RANS Volume of Fluid method in modelling slamming

Abstract

The oil and gas industry requires complex subsea infrastructure in order to develop offshore oil and gas fields. Upon installation, these components may encounter high slamming loads, stemming from impact with the water surface. This paper utilises two different numerical methods, the mesh-free Smoothed Particle Hydrodynamics (SPH) approach and Reynolds Averaged Navier–Stokes (RANS) Volume of Fluid (VOF) method to quantify these loads on a free-falling object. The investigation is also interested in conducting a parameter study and determining the effect of varying simulation parameters on the prediction of slamming event kinematics and forces. The surface impact of a freefalling wedge was introduced as a case study and has been simulated using SPH and RANS, with the results being compared to an experimental investigation. It was found from the SPH simulations that particle resolution and the size of the SPH particle kernel are very important, whilst the diffusion term does not play an important role. The latter is due to the very transient nature of slamming events, which do not allow sufficient time for diffusion in the fluid domain. For the RANS simulations, motion of the wedge was achieved using the overset grid technique, whereby varying the discretising time step was found to have a pronounced impact on the accuracy of the captured slamming event. Through analysing the numerical data, one can observe that the RANS results correlate slightly better with the experimental data as opposed to that obtained from the SPH modelling. However, considering the robustness and quick set up of the SPH simulations, both of these two numerical approaches are considered to be promising tools for modelling more complicated slamming problems, including those potentially involving more intricate structures.