A theoretical and experimental study of the aerodynamics of the curved-bladed darrieus vertical axis wind turbine

• Main Author: Read, Simon
• Published: Kingston University 1986
• Subjects: 629.1323; Aerodynamics
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.372388

The aerodynamic performance of the low solidity curved-bladed Darrieus vertical axis wind turbine has been studied both theoretically and experimentally. Initial studies showed the need for an engineering prediction scheme sufficiently accurate to give blade forces as functions of rotational position which did not require excessive computational time. The scheme proposed here develops a suggestion first made by Lapin in 1975, to treat the turbine as two actuator discs, one upwind and the other down-wind. This suggestion, combined with a multiple streamtube approach, the momentum equations in the freestream direction and blade element theory enables the system of equations to be solved. A continuity argument connects the flow between the two discs. The theory does not rely on analytic formulations for aerofoil force coefficients and can therefore use data obtained from experiment, tabulated for a range of Reynolds numbers, thereby including the effects of stall and drag. Comparison with the power coefficients obtained from experiments, using a two-bladed wind tunnel turbine at a Reynolds number of 28,000 (based on free wind speed and blade chord) shows that the theory is accurate enough to detect the effect of dynamic stall. It is also shown that the continuity argument is essential for improved power output predictions, over earlier single actuator disc theories. The new theory also indicates large differences between air speeds on the upwind and downwind sides of the turbine. These were also confirmed by experiment. Comparison of blade force predictions with those obtained using a computationally expensive time-marching discrete vortex theory shows that good estimates are obtained over the normal turbine operating range. At present only a uniform freestream is treated and turbulence is not accounted for.