Ducted wind turbines revisited: A computational study

Ducted Wind Turbines (DWTs) are one of the many concepts that have been proposed to improve the energy extraction from wind in comparison to bare wind turbines. In reviewing the DWT studies, investigations based on the combined use of theoretical, computational, and experimental techniques have been presented. Although indicated in these studies that the power output of wind turbines can be significantly increased by using surrounding ducts, the factors influencing this power increase, like the duct shape, augmentation add-on’s and yawed inflow conditions, need further investigation. These topics have been addressed in this doctoral thesis. The study presents a computational investigation of DWTs, employing two-dimensional and three-dimensional CFD simulations. To this intent, solutions obtained using panel, RANS, URANS and LB-VLES methods are shown. For reliable solution accuracy, verification and validation assessments are performed when possible. Through parametric investigation, it is found that the aerodynamic performance of the DWT can be improved by increasing the duct cross-section camber and a correct choice of turbine thrust force coefficient, whilst maintaining the same duct-exit-area ratio. The aerodynamic performance improvement for a DWT directly corresponds to the dimensionless duct thrust force coefficient. Flow analysis showed that flow separation when detected inside of the duct, reduces the duct thrust force coefficient and ultimately the aerodynamic performance of the DWT model. In an effort to further improve on the aerodynamic performance of the DWT, the effect of multi-element ducts and Gurney flap on the existing DWT models are investigated. The aerodynamic performance improvement with multi-element ducts strongly depends on the installation settings of the secondary duct element with respect to the primary DWT geometry. On the other hand, a Gurney flap retrofitted at the trailing edge of the duct improves the aerodynamic performance of the DWT model by delaying inner duct wall flow separation, thus increasing the mass flow rate at the turbine. Finally, the effects of yawed inflow condition on the aerodynamic and aeroacoustic performance of DWT models are studied in detail. The analysis showed that DWTs can demonstrate yaw insensitivity up to a specific yaw angle. The yaw insensitivity for the DWT model, however, strongly depends on the aerodynamic mutual interaction between the duct and turbine, which changes with the duct geometry, turbine configuration and yaw angle. While assessing the aeroacoustic performance of the DWT models, it is found that the DWT model xiii xiv Summary with highly cambered duct cross-section generates higher broadband noise levels, which results from the turbulent flow structures convecting along the surface of the duct.