CFD evaluation and experimental comparison on the flow around fixed multi-column configurations

Abstract

The flow-induced motion (FIM) of platforms is a vital phenomenon issue to be investigated to help the designers and to decrease the costs related to the mooring and/or riser systems. On this matter, when talking about multi-column configurations, captive tests (also called fixed or stationary tests) can be a methodology worth of examining the vortex detachments, the source for the oscillatory lift, the drag force, and the interference of the wake of the upstream column on the downstream ones. Computational fluid dynamics (CFD) evaluations on the flow around a fixed array of one, three, and four columns with low aspect ratio, $${\text{AR}} = H/L = 1.5$$ (the ratio between the column draft, $$H$$, and the column face dimension, $$L$$) were conducted. Three different shapes of the column sections were tested, namely, circular, square, and diamond. For the multi-column configurations, the spacing ratio tested was $$S/L$$ (where $$S$$ is the distance from center to center of each column). The Reynolds number tested was 100,000. Lift and drag forces were monitored in each column aiming at the integral quantities of the hydrodynamic to compare with the experimental results. Time averages of velocities and vorticity fields were also observed. The calculations were performed with the flow solver StarCCM+; the code solves the multi-phase incompressible Unsteady Reynolds-Averaged Navier–Stokes (URANS) equations. Simulations were carried out with the Improved Delayed Detached Eddy Simulation (IDDES) method. The IDDES estimates the small turbulence scales near the body boundary with the Reynolds-Averaged Navier–Stokes (RANS) model and the large scales with the Large Eddy Simulation (LES) model. For the volume of fluid (VoF), an interface-capturing approach was used. The multi-column arrays presented no regular vortex shedding pattern when compared with the single-column case. The vortex shedding behavior was mainly determined by the column section. For the multi-column configurations, the number of columns, 3 or 4, was not the main issue. CFD calculations compared with the experiments showed good agreement for the circular and square cases and a significant difference for the diamond cases, which need to be further investigated.