Summary

Simulations have been performed for an HQ17 airfoil equipped with Gurney-flaps at a Reynolds number of $Re=10^6$. Unsteady two- and three-dimensional computations show that the Gurney-flap enhances the mean lift coefficient up to $57\%$. Depending on the flap height, small Gurneys increase it disproportionately more than larger ones. A significant augmentation of drag alongside the higher lift is closely coupled to the appearance of unsteady two-dimensional flow structures in the wake. The numerical simulations show dominant structures corresponding to a constant Strouhal number and strongly depending on the flap height, whereas the effect of incidence is minor. As long as the geometry remains 2d, no important differences are visible between two- and three-dimensional computations.

Results that are in satisfying agreement with experiments can be obtained using the LLR $k$-$\omega$ turbulence model as well as by computations based on a Detached Eddy Simulation. In the case of URANS, very regular structures appear in the wake whereas the DES predicts more complex flow structures which are nevertheless dominated by the same vortex shedding mechanism.

Further investigations provide methods to damp the flow structures and to reduce the aerodynamic drag by modifications of the Gurney design. The simulations demonstrate the effect of splitter-plates, vertical bars (stabilizers) and wake-bodies behind the flap that are able to reduce the drag. Slots in the flap that enhance the three-dimensionality of the flow provide a comparably positive effect.

The Gurney-flap modifications presented exhibit a positive effect on the flow structures in the wake and on the mean drag. Though the mean lift of the standard Gurney-flap is slightly lowered, all modifications still offer a significant extra lift compared to the clean airfoil. The Strouhal number is not dramatically affected. The best overall performance is achieved by a configuration with an airfoil-shaped wake body which provides $40\%$ reduction of the Gurney induced drag in the experiments and up to $85\%$ in the simulation. The fine details of the design have a great impact on its performance. Stabilizers prohibit the effect of vortex shedding on the lift, and $c_l\,'$ remains small which cannot be achieved by the splitter-plate. The drag reduction of both methods is however comparable. Three-dimensional slits in the Gurney can reduce the drag by $12\%$ in the present case.

This simple and cost-effective tool to control the flow can be combined with other devices. Its performance has already been demonstrated in several applications [4,5,25,26,27] and the modification presented enable further adoption to flows where low drag plays an important role.