Aeroacoustic Prediction for Vertical Axis Wind Turbines

Sammanfattning: This thesis investigates the aerodynamic and aeroacoustic prediction of vertical axis wind turbines, using computational fluid dynamics simulations. Noise pollution from wind turbines is one of the disadvantages of wind energy, calling for strategies to reduce noise levels. Yet for vertical axis wind turbines in particular, there is insufficient knowledge of how to identify sound sources and mitigate the sound level. The aim of this study is to predict aerodynamic noise, using large eddy simulation and acoustic analogy, so as to better understand the mechanism of sound generation for vertical axis wind turbines. First, the prediction method is validated for a static single blade in stall. This model is able to capture the dominant frequency, but it does not well reproduce the broadband characteristics. Next, the aerodynamic behavior of the 12 kW H-rotor vertical axis wind turbine is studied, whereby the focus is on the importance of properly modeling the strut influence for an accurate prediction of the blade forces. To achieve this, the flow field is solved for three different tip speed ratios. The results show that the struts significantly affect on the force distribution along the blade. The reduction of the blade force is observed to occur not only at the attachment points of the struts, but also over a large area of the blade section in the downwind side where the blade interacts with the wake created in the upwind. Finally, the noise radiated from the vertical axis wind turbine operating at high tip speed ratio is predicted. Measurements are conducted to validate the prediction, with the experimental data representing the broadband noise characteristics dominant at around 800 Hz. The prediction reproduces the sound pressure level observed at a radial distance of 1.4 rotor diameter, with a few decibels difference. However, these discrepancies become more pronounced at double distance, which can be considered to arise due to the effect of the ground reflection being ignored. The simulation furthermore indicates, that the main sound sources are emitted when the blade rotates in the downwind. It is suggested that future work should properly consider the atmospheric turbulence for more accurate predictions.