Numerical and experimental loads analysis on a horizontal-axis wind turbine in yaw

The characterization of wind turbines in yawed conditions is one of the most important topics as regards the latest advances in optimizing power production and mechanical behavior in wind farms. The classical wind turbine control strategy consists in keeping the rotor constantly aligned with wind direction: whereas this approach maximizes the power coefficient of each single turbine, it might not be the best solution when, in a wind farm, upwind turbines generate wakes on downwind ones. Considering this, yawing the rotors gives a steer to wakes, improving the flow on downwind turbines. This new kind of control strategy has been attracting the scientific interest not only by an energetic point of view, but also as regards the mechanical behavior of turbines operating not aligned with wind, in particular for what concerns generation of forces and vibrations. On these grounds, the aim of this paper is to study in deep how a wind turbine works on yawed configurations. In order to do this, wind tunnel tests have been performed with yaw angles that range over from −45◦ to 45◦ on a 2 m. diameter small scale wind turbine. Experimental measurements of forcespower and tower vibrations are then compared with the results of simulations from two different codes. The first, called BEM, is internally developed following the principles of Blade Element Momentum theory and it is used to estimate forces and torque acting on the rotor. The second implemented model is developed using the FAST (Fatigue, Aerodynamics, Structures and Turbulence) software, developed at the National Renewable Energy Laboratory (NREL). FAST simulations provide in output forces, torque and vibrations of tower and blades. Simulations are set up with similar conditions as the wind tunnel tests, with many yaw angles and steady wind speed. One of the main results of this study is that there is a remarkable agreement between simulations and measurements as regards the estimate of the power coefficient CP in yawed and non-yawed configurations. In spite of this, thrust coefficient CT is not faithfully estimated when the yaw angles is vanishing. This matter of fact is then explained by the fact that low-fidelity numerical models are not capable in reproducing reliably the effect of the tower blockage, slowing down the air stream in its proximity. As a consequence, when a blade passes close this area of reduced flow speed, the generation of aerodynamic forces decreases. In yawed configurations, this phenomenon is less relevant because of the increased distance between blades and tower on air flow direction.