Vortex-Induced vibrations (VIV) of a circular pipe or a riser are caused by flow separation and vortex shedding from the riser when exposed to ocean currents. This phenomenon is frequently observed in the field for drilling risers as well as production risers. Such phenomenon is not desirable as bending stresses due to the vibrations can cause significant fatigue damage to the system. Riser fairings and/or strakes are generally used in the field to streamline the flow and eliminate VIV. Accurately predicting riser VIV responses with or without VIV suppression devices is one of the key challenges for designers and researchers in the relevant offshore industry.
In this study, the VIV benchmark study organized by ITTC Ocean Engineering Committee in 2013 was revisited. A commercial RANS-based CFD code was used to predict drag, lift, and Strouhal number of flow past a fixed long circular cylinder at Reynolds numbers close to the drag crisis. Both 2D and 3D cases were studied. Two turbulence models, namely LES and DES were used. Grid and time-step sensitivity analyses were also conducted. Results were compared with the ITTC benchmark data. This study shows that it is often insufficient to consider VIV of a long symmetrical cylinder as two-dimensional system due to three-dimensional nature of vortex structures. In addition, accurate VIV predictions also require flow in the viscous sub-layer to be fully resolved. These requirements lead to excessive number of cells to carry out CFD simulations. In order to determine effective domain size and practical mesh discretization for VIV simulations, this study included an investigation into the effects of length to diameter ratio of the computational domain. It was found that increasing length to diameter ratio to 2 improved drag and lift predictions significantly and the results agreed well with the benchmark data.
Preliminary simulations were also conducted with a circular cylinder fitted with a generic strake to evaluate the VIV characteristics. Results were compared with experimental data where a close match was obtained for both drag and lift. This also revealed the effects of the strake on the VIV of the pipe. For the present cases, major flow features including pressure, velocity, and vorticity fields are also presented. Three-dimensional effects and unsteadiness were well captured in both turbulence models. LES-based turbulence models seem to be the key to better solve and predict the flow problem numerically. However, their considerable computational demand still does not allow applications for engineering design purposes.