CFD for future America's Cup Challenger

Bettar el Moctar, Jochen Marzi, Arndt Schumacher, Hamburg Ship Model Basin (HSVA)

The Hamburg ship model basin (HSVA) was asked to join the expert team and perform the hydrodynamic assessment of the new ship hull. The main activities of HSVA comprised CFD computations and naturally experimental tests in their large towing tank. The CFD analysis was based on potential flow results using the in-house n-SHALLO code and the Comet RANS code from CD-adapco. The use of a parallel Linux-PC-cluster, enabled a large number of computations to be performed for the different sailing conditions that need to be considered to assess the quality of a new design.

Validation experiments have been performed in the large towing tank for the most promising candidates. Here, the sophisticated planar motion carriage (CPMC) is used to control different operational conditions including all trim, heel and yaw variations that need to be modeled. Fig. 1 shows the experimental set-up during measurements.

The numerical grid for the RANS computations consisted of 3,000,000 hexahedral cells. Rudder, keel, bulb and winglets were considered in the computation. The k-W model was used to account for turbulence effects, while free-surface effects were modeled by an interface-capturing (VOF-like) method using the HRIC (high-resolution interface-capturing) scheme. For the momentum equations, convective fluxes were discretized using a central differencing scheme.

The unsteady simulation time step was chosen such that the Courant number of 1.0 was not exceeded.

Fig. 2: Comparison of computed and measured forces for different designs at different sailing conditions

Each simulation was performed using six processors from our Linux-PC cluster. As shown in Fig. 2, the difference between computed and measured forces for different designs (Design 1 and Design 2) at different sailing conditions (0° and 27° heel angle) was, in most cases, less than fivepercent. The RANS computations for the different designs always predicted the same trends as found in experiments. The simulations have thus proven their ability to account for relative changes in resistance when design or sailing conditions are changed, almost as reliably as experiments. This allows a reduction in the number of experiments needed for validation purposes, since these can be limited to the optimum design found by simulation. Fig. 3 shows the pressure distribution on the yacht body, with details of keel and winglets. Fig. 4 presents cross-flow velocity vectors in a cross-section through the keel for a heel angle of 27°, showing a complex flow structure. Fig. 5 shows free-surface deformation around the yacht for the case of 27° heel angle, with an asymmetric wave pattern. Red areas indicate high and blue areas low water level. Finally, Fig. 6 shows the yacht’s wetted surface as seen from the front for the 27° heel angle condition.

These results demonstrate the suitability of RANS in CD-adapco’s Comet solver for the analysis of flow around yachts under sailing conditions as a complement to experimental testing. While in the early design phase, potential flow solvers are still the only viable simulation tool due to their efficiency (a few minutes of computing time) and the possibility of running hundreds of test cases for various design options RANS solvers are needed to accurately account for the effects of turbulence and high free-surface deformation, including breaking waves.

In future, coupled analysis of wind flow around sails and water flow around hull and appendages in sea waves will be performed to model the interaction of all yacht components affecting its performance.

Although the team were not ready for the America's Cup challenge in 2002, the results obtained in the CFD and experimental investigations provide a valuable basis for the future. Now, the cup being back in Europe for the first time since 1851, expectations are high and the Pinta Racing team is hoping to establish a new challenge for the next round, most likely in 2007. We all hope that at that time it will be possible to demonstrate the usefulness of CFD results in race conditions.

Fig. 1: Yacht model during experiments
in the HSVA towing tank



Fig. 3: Pressure distribution on yacht body
including rudder and keel with winglets



Fig. 4: Velocity vectors in a cross-section
through keel normal to sailing direction




Fig. 5: Free-surface deformation around
yacht (Design 1) at 27° heel angle




Fig. 6: Wetted hull area (red) for the yacht hull at 27° heel angle, viewed from front, indicating wave breaking and air entrainment on the left-hand side