This thesis aims to improve our understanding of wind turbine flows and power efficiency using large-eddy simulation (LES), while also testing and refining LES and analytical modeling approaches.

The first part of the thesis, which can be considered as a preliminary part, focuses on developing and validating numerical tools for wind turbine simulations using LES. First, a coupled LES and blade element momentum (BEM) framework is introduced to characterize the aerodynamic performance of the WiRE-01 miniature wind turbine airfoil, with particular attention to the impact of inflow turbulence. The study provides the airfoil data necessary for further studies, and the validation against experimental data demonstrates the framework accuracy and computational efficiency. Second, a study validating LES with actuator disk models (ADM) for wind turbine simulations is conducted. Several details of the implementation are evaluated based on a test case studied experimentally. Validation with wind tunnel data confirms the method accuracy for wake flow and power predictions, and guidelines are provided for optimized implementations. Together, the two studies provide a complete LES-LES approach for wind turbine simulations.

The second part of the thesis investigates the influence of the base flow characteristics on turbine flows and power efficiency, and their connection. First, the typical scenario of a turbine located at the top of a hill is studied. In particular, the effect of the hill geometry and turbine position are systematically studied. The results demonstrate that hill-induced base flow acceleration/deceleration and associated pressure gradients are critical to the turbine wake and power efficiency. Specifically, adverse base flow pressure gradients in the turbine wake region reduce wake recovery and power efficiency. A key result is the strong link between the turbine wake and power efficiency: the base flow characteristics modify the wake, which in turn alters the flow through the turbine and ultimately affects its power efficiency. In this context, a robust linear relationship is identified between the turbine induction factor and the power coefficient. Second, the role of incoming turbulence intensity is analyzed, exploring the same connection between the turbine wake and the turbine power efficiency. Higher turbulence levels are shown to enhance turbine power efficiency by promoting faster wake recovery, and a robust linear relationship is identified between the turbulence intensity and the power coefficient.

The third part of the thesis focuses on developing analytical models to predict the effects of topography on turbine wakes and power efficiency. First, the classical 1D momentum theory is extended to incorporate pressure gradient effects on the power predictions. This model is validated on idealized ramps using large-eddy simulation and experimental data, demonstrating good accuracy. Finally, a comprehensive analytical framework is proposed to model turbine wake deficit and power efficiency in hilly terrains. The framework integrates pressure gradient and turbulence intensity effects into both wake deficit and power efficiency predictions. Compared to classical approaches, the framework significantly enhances prediction accuracy, highlighting the critical role of the effects of pressure gradient and turbulence intensity induced by the terrain.