Airfoil Design for Vertical Axis Wind Turbines

Abstract

In the past, the vertical axis wind turbine (VAWT) lost the competition to its horizontal axis counterpart, having (wrongfully) presumed disadvantages like increased fatigue and low efficiency. Little research and development into the VAWT led to a large gap between the maturity of technology of both turbine types. Now that due to modern techniques the VAWT experiences renewed interest, topics like airfoil design are reinvestigated. One phenomenon particularly, plays an important role in VAWT aerodynamics: flow curvature. Due to the VAWT airfoil’s rotation about an offset centre, there exists a chordwise angle of attack variation. Also, the added rotational velocities will alter the pressure distribution and boundary layer of the airfoil, leading to a change in force distribution and turbine efficiency. This research sets out to chart these flow curvature effects and implement suggested models to cope with them. Furthermore, using a pitching airfoil model, an implementation in the airfoil analysis code XFOIL will provide the possibility to investigate inviscid and viscous flow curvature effects. What effect these have on airfoil design shall be found using an airfoil optimiser, generating optimal airfoils for various levels of flow curvature. This thesis has applied six models authors suggested to investigate flow curvature. By applying transformations to virtual airfoils, leading to added camber and incidence angle, the researchers thought to mimic rotating VAWT airfoils in straight flow. Comparison with each other and a solution of a verified panel method showed that their approximation is reasonable within bounds. When the pitch rate and location is kept to practical values, also the difference between the methods is very little. Therefore the methods can be used interchangeably. A pitching model has been devised, showing that the chord-to-radius ratio an be simulated by the pitch rate of a pitching airfoil. As the chord-to-radius ratio determines the amount of variation of angle of attack over the chord due to flow curvature, this model can be used to investigate flow curvature effects. The model has been applied in XFOIL and the results show a good comparison to the benchmark panel code. There still exists a discrepancy between the solutions, inherent to the difference calculation methods and dependent on the pitch location. The latter determines the arm to the surface and therefore the magnitude of rotational velocity, which increases the error with respect to the benchmark. Cambered airfoils and airfoils at a larger pitch setting showed less susceptibility to flow curvature. The modified version of XFOIL is used in an optimisation campaign to investigate the impact of increasing chord-to-radius ratio. The method proved to be slightly unstable and therefore little data is obtained. Of the dataset the following trends can be obtained, using the least-squares method. With increasing chord-to-radius ratio, the camber tends to become more negative and shift forward. This is the optimiser trying to maintain optimal power, by choosing airfoils which will stall later to avoid power loss. Also, this results in thinner profiles with their maximum thickness more aft. Finally, the optimal airfoils have been applied in a two-dimensional VAWT analysis model. This showed that there is very little power gain, and possible power loss when applying the pitch optimised airfoils.