Numerical Modelling of Steady and Unsteady Sail Aerodynamics
Examensarbete för masterexamen
Naval architecture and ocean engineering (MPNAV), MSc
A method for prediction of aerodynamic forces produced by a jib and main sail plan for steady and unsteady upwind conditions has been developed, and validated against experimental results. Computational fluid dynamics, with a RANS approach, is used to model the viscous ow on a stiff, model scale sail plan, where the sail shape has been captured in the experimental studies used for validation. An unstructured, overset grid is used, allowing movement of the sail plan. Two rounds of simulations were performed, where the second was started since the tunnel geometry as well as experimental results used for the first round of simulations proved to be incorrect. Correct tunnel geometry and experimental results were later provided by one of the authors of the experimental paper. For the initial simulations, a verification and validation was performed for steady conditions. The verification was successful, and the solution shown to have insignificant grid dependence for the tested grids. However, due to undocumented corrections applied to the experimental data, the validation was unsuccessful, showing large differences. This prompted an investigation concerning physical modelling, where the effects of tunnel conditions, turbulence modelling, scale effects as well as grid topology were investigated. For the final simulations, the correct tunnel geometry was used, and results of the physical modelling investigation were considered. Again, verification and validation were performed. In contrast to the verification in the initial round of simulations, some grid dependence is apparent for the tested range of grids. For the steady case, the subsequent validation was successful, with a 2:8% difference for the driving force coefficient Cx and a 3:2% difference for the side force coefficient Cy. The method was also validated for one unsteady, pitching case. A low amplitude, high frequency pitch motion was tested, and showed good agreement for mean values, with a 2:1% difference for the driving force Fx and a 3:6% difference for side force Cy. The phases of the forces were well predicted but force amplitudes were under predicted, with a 14:3% difference for the amplitude of Fx, and a 18:5% difference for the amplitude of Fy. The prediction of mean value and amplitude for vertical centre of effort zCE was poor, with a 10:2% difference for the mean value, and a 52% difference in amplitude.
Farkostteknik , Transport , Vehicle Engineering , Transport