Kite-like wind energy generation systems can power a generator on the ground.
This innovation can control the flight of a tethered vehicle, in an airborne wind energy (AWE) generation system, through the use of a pan-tilt platform and a visible spectrum digital camera, combined with tracking and control software running on a standard PC.
One of the main technical challenges involved in building a feasible AWE system is developing the control system. It is not sufficient to stabilize the kite — it must be flown in a high-speed crosswind trajectory for power generation. Additionally, the behavior of the kite is highly nonlinear and difficult to model due to flexible structures that assume a shape in response to aerodynamic loads. In this invention, all sensors are kept on the ground, allowing the kite to be as light as possible. A fuzzy logic system was chosen to perform this control task because a reliable model of the kite is unavailable, and the operation of the kite is relatively easy to explain. A commercial, two-line-controlled, ram-air kite was used for developing and testing the control system.
The system uses measurements of the tension in each line, the length of each line, and the estimated position of the kite from a vision-based tracking system as input. It then outputs a commanded difference in length between the lines, as well as an average reel rate for both lines.
A predictor-corrector approach is used to track the air vehicle. The previously estimated position and velocity are propagated to a position estimate. Detecting the air vehicle is accomplished using three algorithms: change in time of brightness, a change in time of hue, and spatial maximum of hue. A pan-tilt platform is used to keep the air vehicle within the camera frame. A fuzzy logic-based controller is used to generate steering and reeling commands for the air vehicle from the position and velocity estimated by the tracking system, as well as the line tension, velocity, wind speed, and direction. Ground-based visual tracking is likely to have higher error in localization compared to a combination of onboard sensors, but has a significant advantage in expense, complexity, and the weight of the air vehicle. A fuzzy controller has advantages in robustness over many other control methods.
This innovation makes use of OpenCV code for some vision processing tasks, Microsoft DirectX and XAudio2 code for communicating with some computer hardware, proprietary libraries for communicating with weather stations, and tension-measuring hardware.
This work was done by David North and Mark Aull for Langley Research Center. LAR-18246-1