CFD analysis of vortex - induced motions of bare and straked cylinders in currents

It is widely acknowledged that the use of helical strakes for mitigation of vortex-induced motions (VIM) of surface piercing cylinders, such as spar platforms, is only partially effective. Using computational fluid dynamics tools, we compare the oscillation characteristics of a bare cylinder and a straked cylinder in uniform currents. Our model comprised of a straked cylinder with diameter of 0.741 m, aspect ratio of 1:1.9 and three helical strakes of height 13% of cylinder diameter. This geometry corresponds to the hard tank geometry of a scaled truss spar model known to exhibit VIM in tow tank testing. In the CFD simulations the cylinder is moored with linear springs to provide a range of reduced velocities. The fluid domain is made of an unstructured grid comprising of hexahedral elements. Fluid structure interaction utilizes grid stretching and a user defined function for solving the equations of motion. Turbulence modeling uses Detached Eddy Simulation (DES) and the boundary layer is modeled using a wall function with a surface roughness of 0.0003 m. Reynolds numbers are in the range of 50,000 to 100,000. Results for straked cylinder compares reasonably with published results, but under-predicts the peak response. In comparing with corresponding results for a bare cylinder without strakes, the spectral features of the transverse displacement show variations, which are found to be due to the spoiling effect of the strakes. INTRODUCTION It is well known in the offshore community that cylindrical bluff structures such as spar platforms suffer from vortexinduced motions (VIM) in strong current conditions. * Corresponding author, [email protected] Specifically, the loop current condition in the Gulf of Mexico, where current speeds may exceed 4 knots (2.05 m/s), can create a conducive environment for VIM in spars. All the classic spars in the Gulf of Mexico have recorded VIM in strong loop conditions (Yung et al. 2004). Helical strakes have been used on spars to reduce these motions. These strakes are normally three-stranded, extending over the lower two-thirds of the spar hull. The height is typically 10% of the spar diameter (Yung et al. 2004). The effectiveness of strakes on spars has not been as remarkable as in air, leading one to contemplate if the basic physics of the problem may have been overlooked. The strakes around a cylinder work by disrupting the correlation of vortices along the length, thus reducing the net transverse force on the cylinder induced by the flow. Bearman and Brankovic (2004) argue that the strakes may work by way of introducing threedimensionality in the separated flow and destroying regularity of vortex shedding. Experience with risers and spar platforms indicates that strakes suppress the motion to some extent, at the penalty of considerable increase in drag forces on the structure. The latter can have implications on the mooring loads for a spar platform. Irani and Finn (2005) have conducted model tests on a truss spar platform with different strake configurations. They conclude that the geometry and orientation of strakes can be optimized for maximum performance. Bearman and Brankovic (2004) conducted experiments on cylinders with passive controllers of vortex-induced vibration (VIV). Their cylinder models were of diameter 0.044 m, and length 0.6 m giving an aspect ratio of 1:13.6. The experiments covered a mass ratio of 2.58 and Reynolds numbers in the range 10 – 10. They found that strakes were more effective than bumps on cylinders in reducing the magnitude of vibration. 1 Copyright © 2005 by ASME However, their conclusion was that the performance of strakes in water was not as effective as in air. Their reasoning based on simple theoretical development of Bearman (1984) is explored further in the present paper. Moored structures in the ocean such as truss spars exhibit motions in all six degrees of freedom. However, it is principally the horizontal plane translations (surge and sway) that are influenced by VIM. Jauvtis and Williamson (2004) studied the VIV motions of a smooth cylinder with two degrees of freedom at low Reynolds numbers. Their experimental program involved studying cylinders of diameters ranging from 38 – 50 mm and aspect ratios varying from 1:7 – 1:10. Results showed negligible influence of inline motions on the transverse motions for cylinders with high mass ratios greater than 6. For lower mass ratio cylinders, there was a remarkable increase in transverse VIV amplitudes of up to 1.5 diameters. Accompanying this was a modified flow field around the cylinder made up of three vortices for ever half cycle, which the authors labeled as the “2T” mode of vortex shedding. We study the vortex-induced motions of bare cylinders and straked cylinders using computational fluid dynamics (CFD) tools. The cylindrical structures used here resemble the hard tank of a truss spar model described by Irani and Finn (2005). A companion paper by Halkyard et al. (2005) provides the benchmark study of CFD with model experiments, as well as examines the influence of current direction on the VIM of straked cylinders. THEORETICAL DEVELOPMENT Sumer and Fredsoe (1997) show through dimensional analysis that the governing non-dimensional parameters for VIV of a long flexibly mounted cylinder of diameter D in a current of velocity U are: Vrn, M, Ks, Re and ks/D For three dimensional cylinders such as spars, one logical extension will be to add the aspect ratio (D/h; h – cylinder draft) as an additional parameter. Here, the mass ratio M is defined as the ratio of the total mass of the cylinder to the displaced mass.