Abstract
Recent studies suggest that vertical axis wind turbines (VAWTs) hold a great prospect of applications in urban and offshore areas. Understanding turbines’ wake aerodynamics, which are characterized by velocity deficits and upsurges in turbulence levels, is of critical importance for the optimal placement of multiple turbines. This thesis reports an experimental and numerical investigation of the wake characteristics of a VAWT with straight cambered blades and the application of an analytical wake model to wind farm layout design.Systematic measurements of the near and far wake ranging from 1-10 turbine diameters were conducted in a wind tunnel. The results showed that the VAWT’s wake demonstrated significant asymmetry in the horizontal direction, whereas it was roughly symmetrical against the blade mid-span plane. In addition, counter-rotating (CR) vortices evolving in the wake were identified. Integral length scales indicated that the vortices grew along the downstream distance. To further examine the complex wake aerodynamics, a full-scale three-dimensional (3-D) computational fluid dynamics (CFD) model with a structured mesh was built. Independence checks on the spatial discretization, temporal discretization, computational time, and size of the rotational zone were conducted. The algebraic wall-modeled large eddy simulation (WM-LES) was chosen as the subgrid scale (SGS) model. The LES model successfully reproduced the wind tunnel test (WTT) data. The LES results confirmed the wake asymmetry. Moreover, the vortex-ring (VR) structures were identified by the LES model with the Q criterion. The results showed that the counter-rotating vortices existed not only at the windward but also at the leeward.
In addition, the effects of grid turbulence on wake characteristics, self-starting, dynamic stall, and power performance were examined by both the WTTs and LES. Turbulence was generated by installing a wooden grid upstream of the VAWT. The WTT results revealed that turbulence was beneficial to the VAWT’s self-starting and wake recovery. The wake immersed in turbulence demonstrated stronger asymmetry than that in the smooth flows. In both flow regimes, the LES-predicted velocities agreed well with the WTT data. In addition to better self-starting behavior, the LES predicted that operating in turbulence greatly improved the VAWT’s power performance. The turbulence helped to energize the boundary layer flows and hence delayed the occurrence of dynamic stall. Moreover, the VR structures in the grid-turbulence flows were highly distorted.
Finally, an analytical wake model was developed based on the measurements in the WTTs and comprehensive information about the flow fields gained from the LES. To take into account the wake asymmetry in the horizontal direction, two semi-ellipses separating the wake from the free-stream flows were proposed. The wake edges represented by the semi-ellipses were assumed to linearly spread outward downstream. The velocity in the wake was derived according to the conservation of continuity. The parameters of the wake model were derived from the experimental data. Based on the analysis of the wake asymmetry, an approach was proposed to determine the position of a downstream VAWT relative to the wake of its upstream counterpart. The micro-siting of multiple turbines was eventually converted into a mathematical optimization problem. Furthermore, the layout design was successfully generalized to wind farm sites with non-rectangular boundaries. Applications to wind farm scenarios suggested that the resulting layout designs effectively increased total power production. This clearly demonstrated the significance and necessity of wind farm layout design.
| Date of Award | 22 Jul 2016 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Heung Fai LAM (Supervisor) |