Assessment of hydrodynamic performance of a ship hull has been focused on a model ship rather than a full-scale ship. In order to design the propeller of a ship, model-scale wake is often extended to full-scale based upon an empirical method or designer's experience, since wake measurement data for a full-scale ship is very rare. Recently modern CFD tools made some success in reproducing wake field of a model ship, which implicates that there are some possibilities of the accurate prediction of full-scale wakes. In this paper firstly the evaluation of model-scale wake obtained by Fluent package was performed. It was found that CFD calculation with the Reynolds-stress model (RSM) provided much better agreement with wake measurement in the towing tank than with the realizable k-ε model (RKE). In the next full-scale wake was calculated using the same package to find out the difference between model and full-scale wakes. Three hull forms of KLNG, KCS, KVLCC2 having measurement data open for the public, were chosen for the comparison of resistance, form factor, and propeller plane wake between model ships and full-scale ships.

SUMMARYTo enhance the applicability of CFD techniques to the hull form design of modern commercial ships, an e cient and robust numerical method for turbulent ow calculation is developed. The preprocessor is composed of hull form presentation, surface mesh generation, and ÿeld grid generation. The ÿnite-volume method is utilised to solve the Reynolds-averaged Navier-Stokes equations. Three k-turbulence models, i.e., the standard k-model (SKE), the RNG-based k-model (RNG), and the realisable kmodel (RKE), are evaluated to investigate the di erence caused by the turbulence model. The developed numerical method is applied to two practical VLCC hull forms with the same forebody and the slightly di erent afterbodies in order to ÿnd out whether CFD can capture the di erence of stern ow due to hull form variation. It is found that RKE successfully predict the strength and location of bilge vortex, while SKE and RNG fail. However, all three models can provide the right information on the nominal wake di erence between two hull forms.

The flow characteristics in the stern and near-wake region of two ship models, the Korea Research Institute of Ships and Ocean Engineering (KRISO) 3,600 TEU containership (KCS) and the KRISO 300K very large crude oil carrier (VLCC) (KVLCC), were investigated experimentally. The double-deck ship models were installed in a subsonic wind tunnel. The freestream velocity was fixed at Uo = 25 m/s, and the corresponding Reynolds numbers based on the model length (Lpp) were about 3.3x 106 and 4.6x 106for the KCS and KVLCC models, respectively. The spatial distributions of mean velocity components and turbulence statistics, including turbulence intensities, Reynolds shear stresses, and turbulent kinetic energy, were measured using a hot-wire anemometer. For both ship models, the stern flow and near-wake show very complicated three-dimensional flow patterns. The longitudinal vortices formed in the stern region dominantly influence the flow structure in the near-wake region. In the region of main longitudinal vortices, the mean velocity deficits and all turbulence statistics have large values, compared with the surrounding flow. As the flow moves downstream, the turbulence statistics increase and have maximum values at the after-perpendicular (AP) plane and then decrease gradually due to the expansion of the shear layer. For the KVLCC model, the spatial distributions of mean velocity components and turbulence intensities behind the propeller plane clearly show hook-shaped contours. These experimental results, especially the turbulence statistics, can be used not only to understand the flows around modern practical hull forms but also to validate the computational fluid dynamics codes and turbulence models. The complete experimental data set is available on the website (http://www.postech.ac.kr/me/efml/data).

No abstract

No abstract