In-depth study of dynamic stall control : fundamentals to application

Abstract

Wind turbine blades are growing at remarkable rates as manufacturers aim to increase power capture. Currently, wind turbines that consist of 100 meter long blades are already in the prototype stage, and this trend is expected to continue. As wind turbine blades get larger, the structural dynamics of a long and slender blade become problematic. Specifically, unsteady fatigue loads arise in the inherent rotational nature of the rotor, and can detrimentally impact the fatigue life of the wind turbine blades and its various components. Throughout their lives, wind turbines may experience numerous aerodynamic events induced large structural oscillations; events that can only be controlled by the use of active flow control devices. Specifically, the current work controls structural oscillations that arise due to dynamic stall, using synthetic jet actuators.

The current work postulates the following regarding the control of dynamic stall using synthetic jet actuators: (1) the aggressive loads experienced during dynamic stall are directly related to the vorticity generated and shed from the wind tunnel model, and (2) by pulse modulating the synthetic jet actuators at the natural shedding frequency of the separated flow over the airfoil, the shedding of vorticity, and thus the loads experienced by the wind turbine blade, could be controlled more effectively as compared to the continuously actuated case. The investigation of this technique, referred to as pulse modulation, was a main thrust of the current work.

This dissertation dissects the control of dynamic stall by investigating the various components that contribute to it. A wind tunnel model based on the National Renewable Energy Laboratory's 3-bladed Controls Advanced Wind Turbine was tested experimentally in various conditions, utilizing surface pressure, aerodynamic load, and flowfield measurements. During dynamic stall, the angle of attack of the wind turbine blade is continuously varying in time, but despite this, much of the current work involves experiments at static angles of attack. In these experiments, numerous conclusions were made regarding the performance and control of the wind tunnel model. To support the first postulation, an experiment was conducted to force a separated (stalled) flowfield to reattach using synthetic jet actuators. As the flowfield reattached the lift of the airfoil was increased by as much as 20%. It was shown that as the flow was reattached, multiple vortices were shed in the process, very similar to those shed during dynamic stall. This showed that vorticity flux is a governing factor in the dynamic stall loads. Then, an experiment to support the second postulation was developed. The flowfields were measured when pulse modulated waveform was used to drive the synthetic jets. It was shown that pulse modulation has the ability to strengthen the naturally occurring vortices associated with a separated flow. This increased the lift over even the continuously actuated (not pulsed) technique at certain angles of attack, while saving 65% of the power consumption. More importantly, it was found that downstream of the synthetic jet actuators, pulse modulation shed {more vorticity compared to the continuously actuated method.

It was then shown that a closed loop, active flow control system can be designed, built, and field tested on a utility scale wind turbine with short developmental time frames and limited resources. This was conducted on a 550 kilowatt wind turbine at the National Renewable Energy Laboratory. A fully contained system was used to measure strain loads on the blade and actuate flow control actuators. Loads were measured both in open loop and closed loop configuration and were shown to be reduced when comparing actuated to non-actuated configurations by as much as 7%, indicating that active flow control actuators hold significant potential if integrated into the entire blade design. Overall, the current work has demonstrated that the installation of closed loop active flow control on wind turbines is feasible, and has the potential to provide significant decreases to the cost of energy if properly implemented.

Then, experiments measuring the loads associated with dynamic stall were conducted. Lift overshoot and deviation in the pitching moment coefficient due to the formation of the dynamic stall vortex were shown, in addition to significant hysteresis in the loads (different loads whether the model was pitching upward or downward) due to the inception of stall and the motion of the body. The mechanisms by which dynamic stall occurs, in both shallow and deep stall, were described using volumetric flowfields for the finite span model. Synthetic jet actuators were then shown to be able to improve significantly the detrimental aspects of all dynamic stall conditions. For the shallow dynamic stall scenario, the synthetic jet actuators were able to maintain attached flow near the leading edge throughout the entire dynamic pitching cycle. By maintaining attached flow near the leading edge, dynamic stall was inhibited from forming, eliminating the dynamic stall vortex-induced lift overshoot, reducing the lift hysteresis, and eliminating any excursions in the pitching moment coefficient. When deep dynamic stall was investigated, it was shown that actuation of the synthetic jet actuator was not able to maintain attached flow near the leading edge throughout the entirety of the pitch cycle; however, it was shown that the dynamic stall vortex-induced lift overshoot was eliminated, the lift hysteresis was reduced, the range of pitching moment coefficient loads was reduced, and the structural oscillations associated with the pitching moment coefficient were also notably reduced. This suggests that despite whether the synthetic jet actuators can inhibit flow separation near the leading edge, they still can alter the flowfield and control the unsteady loading.

Regarding the two proposed postulations, it was shown that pulse modulation can improve many aspects of dynamic stall control with synthetic jet actuators, and did so with only 35% of the power consumption. For the shallow dynamic stall case, the hysteresis in the lift coefficient was dramatically reduced when compared to the continuously actuated case. This meant that for a wind turbine blade, the flapwise force was nearly identical whether the blade pitched upward or downward. For the deep dynamic stall case, similar reduction in lift coefficient hysteresis was also observed. The mechanism by which the flowfield undergoes dynamic pitching was presented. The flowfields for the deep dynamic stall case were analyzed and showed a dramatic reduction in the size of the wake as compared to the continuously actuated case with flow separation. Analysis of the change in circulation through the wake of the airfoil suggested that pulse modulation can minimize excursions by reducing the circulation shed from the pressure side during stall inception, and increasing the circulation shed during flow reattachment. The benefits obtained by utilizing pulse modulation, which are reduced lift hysteresis and 65% reduction in power consumption (compared to the continuously actuated case), suggest pulse modulation may have a valuable place in the use of synthetic jets to control dynamic stall.