Automation and Parametric Investigation of Wind Turbine Wakes in an Aero-Servo-Elastic Large Eddy Simulation Framework

Abstract

Wind turbine wakes have been a topic of intense research since the maturation of wind energy. This is justified given the complexity of the physics involved and their crucial impact on the operation of a wind turbine. Only recently though, the advancement in computational resources and the development of models like the Actuator Line Method (ALM), has made it possible to study wind turbine wakes while capturing their multi-scale and multi-physical aspects. In this regard, the Actuator Line - Large Eddy simulation framework has become an academic and industrial standard. Thus, this thesis has been carried out under the joint supervision of Siemens Gamesa Renewable Energy and TU Delft to study wind turbine wake interaction using the coupling between the LES library YALES2 and the servo-structural solver BHawC. The latter extends the framework to include the effects of control and structural deformation. It allows for the simulation of real wind turbines with their industrial controller.

The aim of the project was twofold, involving the development of an automatic workflow for coupled ALM-LES of wind turbine wakes which is then used to gather insights into the impact of inflow conditions on wake properties as well as the structural impact of different partial-wake incidence scenarios.

The workflow integrates the steps of external flow convergence, mesh refinement in the wake region and generation of the converged and time-averaged flow field. It allows parametric studies to be carried out with minimal intervention from the user, while ensuring the reliability of the results. With respect to the external flow, the recycling method was used to obtain fully developed turbulence with sufficient control on the turbulence properties.

An investigation of the inflow conditions showed the sensitivity of the velocity field in the wake to the ambient turbulence and wind speed. An increase in ambient TI from 5\% to 10\% led to 35\% more wake expansion in the lateral direction and 30\% faster wake recovery. The ambient wind speed played an important role in recovery by determining the operating condition of the upstream turbines which in turn affected the wake-added turbulence. An analysis of wake meandering showed that this phenomena is largely dependent on the ambient turbulence.

Lastly, the structural impact of two different partial-wake interaction scenarios was studied to highlight the importance of accounting for the wake position on the rotor of downstream turbines while carrying out load assessments. In one scenario, the turbine operated in half-wake and half free-stream while in the other, it operated under two half-wakes. It was observed that at above-rated wind speed, the first scenario led to a 35\% increase in the flap-wise damage accumulated by the blade. On the other hand, the edge-wise damage changed by 4\%. These results emphasized the need for considering the spatial distribution of the wake on the downstream turbines.