STAR European Conference 2010 London, 22-23 March FREE MOTION SIMULATION OF A SAILING YACHT IN UP-WIND CONDITION WITH ROUGH SEA G. Lombardi, M. Maganzi, A. Mariotti Dept. of Aerospace Engineering of Pisa Italy 1
America s Cup THE OLDEST SPORT EVENT IN THE WORLD IT REPRESENTS THE TOP LEVEL IN SAILING RACES SAILORS TEAM ORGANIZATION TECHNOLOGIES DESIGN High impact To fully develop and promote a green revolution on the America's Cup Competition. Driven by Science and i-technology RESEARCH and DEVELOPMENT Cooperation with University of California, Los Angeles Politecnico Turin University of Pisa 2
Fluid Dynamics In Yacth Design In Yacht Design Fluid Dynamics assumes a crucial role GREEN COMM Challenge planned both experimental and numerical approaches, but the design will be driven essentially by CFD For CFD analysis different levels are foreseen: Analysis on the sails in reference conditions Analysis on the hull in reference conditions Optimisation procedures for several geometrical elements These aspects are well established from previous analyses on different problems (boats, airplanes, cars), BUT: 3
THE PROBLEM These analysis are carried out on simplified configurations and in reference conditions. From the design point of view, it is particularly important the evaluation of the performance in off-design conditions. Interest to set-up a CFD procedure for the performance evaluation of COMPLETE configurations in GENERIC sea conditions STAR CCM+ has the capability to evaluate the motion of the complete boat with rough sea 4
SOFTWARE CAD GEOMETRY CATIA V5 R19 Surface GRID ANSA 13.0.2 Volume GRID and CFD STAR CCM+ v4.04.011 5
HARDWARE Linux cluster: 16 SUN Fire X4100. Each server: 2 AMD Opteron 285 (Dual Core) processors and 4GB RAM each (64 processes) 6
The GEOMETRY America s Cup v5 yacht (2007), in up-wind condition. mainsail mast helm wing bulb jib hull
SURFACE GRID Base size for hull and sails 0.15 m Base size for domain external surfaces 1.20 m Total surface elements 200000 Skewness<0.5. COARSE GRID 8
VOLUME GRID Computational domain: Lenght 100 m Width air height water height 100 m 70 m 20 m Trimmed volume grid with prism layers around the sails Grid refinement at the free surface in order to have a corrected representation of the wave profile. Number of cells in the prism layer 5 Height of the prism layer Total volume elements 0.15 m 16 Mil. 9
Free Surface Refinement Refinement for pitch motion Refinement for roll motion 10
PHYSICAL MODEL 3D unsteady k-ε standard turbulence model Wave model (VOF Waves) 6 degrees of freedom rigid body motion (6 DOF Motion) 11
WAVE MODEL (VOF Waves) Volume of Fluid Waves Model, with two phases (Multiphase Mixture). The sea wave dynamics is represented by the 5 order Stokes theory. The following characteristics are assigned: Wave height, H Wave period, T Sea depth, d Direction and velociy of the sea tide, c E H c E d 12
VOF WAVES set-up Wave direction 35 from the hull simmetry plane Wave height 0.8 m Wave period 3.7 s Sea depth 100 m Sea Tide 0 The corresponding Wave Lenght results l=21.7 m 13
6 DOF Motion set-up Total mass, C.G. position and Inertial Moments are assigned THE MOTION OF THE BODY IS EVALUATED AS A FUNCTION OF THE AERODYNAMICS FORCES ON THE SAILS AND THE HYDRODYNAMICS FORCES ON THE IMMERSED HULL, EVALUATED AT EACH TIME STEP, TAKING INTO ACCOUNT THE INERTIAL CHARACTERISTICS. 5 D.O.F.: x, y, z translations Roll Pitch The yawn rotation is kept fixed to take into account the control carried out by the wheelsman 14
EVALUATION EXAMPLE Wind speed 10 m/s Wind direction 35 from the hull simmetry plane TRANSIENT PHASE: Time Step 0.025 s Iterations per Step 15 Total Time 5 s ANALYSED PHASE: Time Step 0.025 s Iterations per Step 15 Total Time 20 s 15
RESULTS - the BOAT MOTION 16
COMPUTATIONAL TIME TOTAL TIME FOR THE PRESENTED SIMULATION: 1 Month!! Despite the coarse grid, the computational time is very high TO OBTAIN ACCURATE RESULTS IN REASONABLE TIME IT IS NECESSARY TO HAVE A VERY LARGE COMPUTING CLUSTER 17
RESULTS A large amount of data is obtained by the evaluation Information are available on: Boat speed Boat motion Aerodynamics behaviour of the sails Hydrodinamics behaviour of any element of the hull Correlation between different data. Ex.: boat speed and roll angle boat speed and pitch angle pitch rate and lift on the sails EXAMPLES 18
Boat Speed 19
Boat Heel 20
Speed Heel correlation Poor correlation appears 21
Hull Drag 22
Speed Hull Drag correlation More evident correlation appears (with phase angle) 23
SAILS THRUST The Thrust is mainly from the jib There is a phase angle between mainsail and jib thrusts 24
SAIL FLOW REPRESENTATION 25
SAIL FLOW REPRESENTATION 26
FLOW DETAILS. 27
Vorticity Field on the deck 28
CONCLUSIONS The motion of the boat is evaluated as a function of the aerodynamics forces on the sails and the hydrodynamics forces on the immersed hull, taking into account the inertial characteristics. Rough sea is considered The case setting in STAR CCM+ did not shown particularly problems The results appear congruent with the real behaviour of the yacht The accuracy could be increased with a larger computational domain and a refined grid Despite the coarse grid, the computational time is very high To obtain accurate results in reasonable time it appears necessary to have a very large computing cluster 29
ACKNOWLEDGEMENTS Thanks are due to A. Ciampa and E. Mazzoni (INFN of Pisa) to render the computing system very efficient and easy to use, resulting from their research activities on computing networks, applied to our cluster.
THANK YOU 31