Abstract
Wind tunnel studies using model wind turbines have been used to investigate the effects and characteristics of neutral and unstable atmospheric boundary layers on their operation and wake behaviour. Wind turbine arrays have also been arranged to observe the effect of wake interaction. Single-point two-component and two-point single-component velocity measurements have been made using laser Doppler anemometry in conjunction with cold-wire anemometry to interrogate the modelled boundary layer. The manufacture and installation of a second traverse mechanism in the wind tunnel was necessary to perform the two-point measurements, along with the development of laboratory software for control and data analysis. In order to allow for measurements of turbine performance, a current sensor was developed so that correlations could be made between velocity and torque fluctuations. Investigation of larger arrays, up to 12 turbines, required the production of additional turbines and installation and subsequent integration of the associated control systems. Measurements made in the neutral flow conditions show that there is an increasing correlation between the upstream turbulence and torque fluctuations with proximity to the turbine, especially in the wake of another turbine where the flow is rapidly evolving. Two-point velocity measurements, with a lateral separation, have shown that there is little effect of the turbine on the correlation of the flow over the rotor disc. Analysis of data from this type of measurement also shows that in an array of four aligned turbines, the spatial structures reach an equilibrium state and are of larger size after the second turbine. Furthermore, the velocity-torque correlation magnitude decreases after the first turbine, but then increases with distance through the array owing to the increased correlation over the rotor disc, although not to the level observed for the first turbine. The turbulence approaching the first turbine behaves in a frozen-flow manner, but this is not true for the second and subsequent turbines, although the idea of convection time still applies. Measurements made in the modelled unstable atmospheric boundary layer show that the length and time scales are changed in the flow, in addition to the alteration of the profiles of mean velocity and Reynolds stresses. The increased turbulence caused by the convective boundary layer increases the rate of wake deficit recovery and does not result in the same spatial structures as the neutral conditions. Temperature effects are of secondary importance with regard to wake and turbine behaviour, with the main driving force behind the performance being the increased turbulence levels.