Simulation methods of wind turbine aerodynamics currently in use mainly fall into two categories: the first is the group of traditional low-fidelity engineering models and the second is the group of computationally expensive CFD methods based on the Navier-Stokes equations. For an engineering environment the search is for “medium fidelity” wind turbine simulation methods that bridge the gap between the computationally inexpensive low-fidelity methods and the computationally expensive CFD methods. The present study focuses on the development of the theory and the practical implementation of a fast multilevel integral transform in a computer program. We utilize this multilevel scheme in a low-order panel method. It is demonstrated that for the simulation of the wake flow of wind turbine rotors the computational burden is reduced from O(N2) for a conventional panel method to O(N) for the present method. This implies that the computational effort is reduced to grow linearly with problem size N, with N the number of panels. As a validation test the panel method, combined with the MLMIC scheme for the influence of the wake, is applied to the MEXICO wind tunnel experiment. The numerical results show good agreement with the experimentally obtained pressure distributions at 5 blade sections of this model wind turbine as well as the results of a state-of-the-art solution method for the Reynolds-averaged Navier-Stokes equations. Compared to a conventional panel method, the MLMIC scheme reduces for the used panel discretization of the three rotor blades and their wake surfaces, the computation time for the wake deformation by a factor 150. With the results achieved in this thesis regarding computational times and simulation accuracy we conclude that the panel method is repositioned as a valuable medium fidelity numerical design tool for wind turbine rotor blades in an industrial engineering environment. The new method thus forms an important step for renewable (wind) energy.
|Award date||14 Dec 2016|
|Place of Publication||Enschede|
|Publication status||Published - 14 Dec 2016|