Dynamic-stall-driven vertical axis wind turbine: An experimental parametric study

Dynamic stall on the blades of low-solidity vertical axis wind turbines (VAWTs) is a major problem due to its adverse effects on performance, drive-train components, and generator sizing. However, when the turbine chord-to-radius ratios are relatively large, greater than approximately 0.5, counterintuitively, dynamic stall can be harnessed to produce useful torque. To examine this seeming contradiction, a wind tunnel study was conducted on a two-bladed, H-rotor test-turbine where the blade chord-to-radius ratio, blade profile (NACA 0012 and NACA 0021), preset angle, strut-blade offset (connection point), and Reynolds number were varied systematically. Large chord-to-radius ratios, between 0.6 and 1.2, and hence high solidities, were selected to ensure that the maximum driving torque was produced by dynamic stall. At low Reynolds numbers (<105), turbine performance with NACA 0021 blades was vastly superior to that with the thinner NACA 0012 blades, producing approximately 90% greater power and torque coefficients at blade tip-speed ratios between 0.8 and 1.6. This difference was due mainly to the vastly different dynamic stall behavior of the two profiles. The high turbine power coefficients, CP,max>0.3, were obtained with the NACA 0021 blades when the preset angles were close to zero and the strut-blade offset was at the mid-chord location. Tuft-based flow visualization showed that an aft dynamic stall vortex on the NACA 0021 blades precedes the conventional leading-edge vortex, also present on the NACA 0012 blades, resulting in substantially greater post-stall dynamic lift. Nevertheless, NACA 0012 showed a far stronger Reynolds number dependence. In addition, the similarity between preset and offset effects was explained by a new approach to describing the virtual preset angle dependence on the offset point.