Aerodynamic Analysis Tool for Dynamic Leading Edge Inflated Kite Models

Abstract

Kite Aerodynamics Airborne Wind Energy (AWE) systems operate at high altitudes, where wind velocities are higher in magnitude and more constant with respect to lower heights at which for example wind turbines operate. The ASSET department of the Delft University of Technology uses a concept which consists of a flexible Leading Edge Inflated (LEI) kite, which is connected through a tether to a drum on a ground station. By autonomously flying the kite in crosswind patterns, high tension forces in the tether cause it to be unwound from the drum, this mechanical energy is converted into electricity via a generator. This kite power system has the potential of becoming a new player in the wind energy sector. The development of such systems creates a demand for kite models, which can be used to optimize the performance of tethered wings. By using aeroelastic models, which simulate the Fluid-Structure-Interaction (FSI), the flight dynamics of a flexible wing can be studied. In this thesis an aerodynamic analysis tool is presented, which can serve as part of a quasi-steady aeroelastic model for kites. The New Tool The Vortex Lattice Method (VLM) and 3D panel method were identified as acceptable aerodynamic modeling methods with the best balance between model complexity and computation time. Because the tethered wings in AWE systems mostly y at high angles of attack, close to stall, an extension is implemented. This extension corrects for the non-linear aerodynamic phenomena that occur at high angles of attack. A non-linear potential ow solver is used, where viscous corrections on both lift and drag are implemented by using airfoil data of the analyzed wing. The method combines the capability of a VLM and 3D panel method to incorporate effects of finite, non-planar wings with the capability of viscous airfoil analysis to predict non-linear effects, including stall. The method is implemented in an already existing analysis tool for airfoils, wings and planes, named XFLR5 which is programmed in C++. Several approaches within the method are considered where, with a trade-o_ validation procedure, the method with the best performance is selected. Because the applied corrections in the method largely depend on viscous airfoil data, an investigation is performed on the single membrane airfoils used in LEI kites. Results are compared for high-fidelity 2D CFD simulations, the 2D integral boundary layer panel method XFOIL and an existing polynomial regression model based on single membrane airfoil CFD simulations. The polynomial regression model is considered most promising, though improvements on the current available model could significantly increase accuracy of the non-linear methods. For the chordwise pressure distribution, an approach is given where XFOIL is used on-the-fly.