Identification of wind turbine testing practices and investigation of the performance benefits of closely-spaced lateral wind farm configurations

The current thesis investigates the development of guidelines for testing small-scale wind turbines and identifies a method that can be used to increase the performance of wind farms. The research was conducted using two scaled wind turbine designs. The first design was a three-bladed wind turbine designed in the Department of Mechanical and Aerospace Engineering (MAAE) to operate in a low Reynolds number regime and to generate a thrust coefficient representative of commercial-scale wind turbines. An Eppler E387 airfoil was selected for the wind turbine due to its behaviour at low Reynolds numbers and the chord of the turbine was increased relative to full-scale designs in order to increase the range of Reynolds numbers that could be attained. The second design was a geometrically-scaled version of an existing two-bladed wind turbine with a NACA 0012 airfoil that was originally designed at the Delft University of Technology.;Following the identification of testing practices related to blockage, the effect of the Reynolds number on the development of the initial wake expansion was investigated using two different rotors. The wake expansion downstream of a 25 cm diameter, three-bladed MAAE wind turbine became less sensitive to the Reynolds number above a Reynolds number of 20,000. This behaviour may be related to the laminar-to-turbulent transition behaviour of the E387 airfoil on the rotor blades. The wake downstream of the geometrically-scaled rotor was found to be 40% to 60% narrower than the initial wake expansion downstream of the corresponding medium-scale rotor. The work identified the need to develop a wind turbine design for a particular Reynolds number regime as opposed to merely geometrically-scaling a turbine.;The performance of scaled wind farm configurations was then evaluated using 20 cm diameter MAAE wind turbines installed in the 1.68 m × 1.12 m atmospheric boundary layer wind tunnel at Carleton University. A scaled boundary layer was generated using triangular boundary layer spires and roughness elements installed along the upstream fetch of the tunnel. Each wind turbine was outfitted with a DC generator and the power output generated by the scaled turbines was used to characterise their performance. A single-normal hot-wire probe was used to determine the mean speed profiles in the fiowfield. Two laterally-aligned wind turbines were separated by a gap and it was observed that when the gap was less than 3 diameters (D), the speed of the flow between the rotors was increased from the rotor plane to approximately 2.5D downstream. This behaviour was identified as an in-field blockage effect and is analogous to the increase in wind speed caused by blockage in a closed test section. The increased flow speed was associated with a narrowing of the wake between the closely-spaced rotors and the concept of capitalising on this in-field blockage effect using a third, offset rotor was investigated. Performance measurements were conducted using 3 gap widths between the outer two turbines and a third, central turbine was placed at 9 different downstream positions. The middle turbine experienced an increase in power when placed within 2.5D of the upstream rotor plane due to the increase in speed in this region. This approach to planning wind farms will help to limit power losses due to downstream wake effects while providing an increase in power output at mean annual wind speeds.;Experiments were conducted in a 0.61 m × 0.81 m water channel in order to independently evaluate the effects of increasing blockage and Reynolds number on the development of the wind turbine wake. Quantitative dye visualisation was used to identify the position of tip vortex cores relative to the blade tip in order to assess how blockage and Reynolds number effects modified the initial expansion in the near wake. Blockage effects on the wake development were assessed using five wind turbines with diameters ranging from 20 cm to 40 cm, corresponding to blockage of 6.3% to 25.4%. The rotors were all operated at a similar tip speed ratio of 6 and a Reynolds number of 23,000 based on the blade tip speed and tip chord. One Outcome of the research was the identification of a limit beyond which blockage narrowed the expansion in the near wake of a wind turbine. It was observed that blockage should be maintained at less than 10% in order to prevent the wake from narrowing artificially due to the flow acceleration around the turbine caused by excessive blockage. The experimental results were compared to a freestream computational simulation of the same turbine using the vortex particle method code GENUVP. The magnitude of the wake expansion in the freestream computation was similar to the experimental wake expansion observed with 6.3% and 9.9% blockage.