The Importance of Scale Effects for Wind Propulsion: Experimental and Numerical Analysis of a Wing Sail

This paper discusses the impact of scaling effects in the performance of a standalone wing sail, comparing experiments and numerical simulations. The experimental data of a standalone wing sail with a NACA0015 section, tested in the R.J. Mitchell wind tunnel at the University of Southampton, are compared with wind tunnel tests run at KTH Royal Institute of Technology, where the L2000 wind tunnel and wing sail had a smaller scale, and with Computational Fluid Dynamics simulations of three different cases. For the numerical simulations, first the model scale Southampton wind tunnel was simulated. Then, keeping the same scale, the wind tunnel domain was substituted with a larger domain to simulate an open-field condition, and analyse the presence of blockage effects. Finally, full-scale simulations were achieved keeping the same scale of the model scale open-field simulations, and reaching the full-scale Reynolds number by varying the viscosity of the fluid. The flow in the numerical simulations is modelled with the RANS equations and the k − ω SST turbulence model, knowing about its limitations in simulating stall conditions, but judged to be satisfactory for a preliminary study about scale effects. The range of model scale Reynolds numbers covered by both experimental campaigns spans from 2.2x105 to 6.7x105, while the full-scale Reynolds number is equal to 7.9x106, covering a range representative of most wind propulsion technologies. The main conclusions are that the simulations capture well the shape of the lift curve up to an angle before stall, after which the simulations diverge from the experiments. In full-scale, higher lift coefficients are reached, and the lift curve shows a different behaviour than in model scale, with a longer linear region and a more abrupt stall.