As wind farms become larger and more closely spaced, predicting their wakes becomes increasingly challenging because their wake evolution depends strongly on turbulence, atmospheric stability, and stratification within the atmospheric boundary layer (ABL) and the free atmosphere. This thesis investigates wind-farm flow dynamics across stable, neutral, and convective boundary layers by combining high-fidelity large-eddy simulations (LES) with physically grounded analytical wake modeling to improve predictive capability while retaining computational efficiency for engineering applications.
First, commonly used analytical wake models are evaluated against LES data for finite-size wind farms of different sizes and layouts under conventionally neutral conditions. These models are found to systematically overestimate wake recovery, particularly far downstream. An improved analytical framework is therefore introduced that relaxes the assumption of linear wake expansion, yielding more accurate predictions of velocity deficits and power output both within and downstream of wind farms.
Second, the physics-based analytical wake model of Vahidi and Porté-Agel (2022) is systematically validated using LES across stable, neutral, and convective ABL regimes. The model accurately predicts wake deficits under all stability conditions and remains robust when Lagrangian integral time scales are replaced by empirical estimates. Unlike classical Gaussian formulations that rely solely on streamwise turbulence intensity, the model explicitly links turbine wake growth to cross-stream turbulence intensities and integral time scales governing lateral and vertical wake meandering and mixing. Building on this validation, a simple stability-dependent modification of the classical Gaussian wake formulation is proposed by relating wake expansion to radial turbulence intensity, thereby improving predictive performance under non-neutral conditions while preserving analytical simplicity.
Third, the combined effects of surface-layer stability and free-atmosphere stratification on wind-farm flow and power production are examined. Increasing stability reduces turbulence intensity, delays wake recovery, and amplifies power losses, with strong stratification enhancing blockage and triggering gravity-wave-induced flow responses that modify induction regions and downstream wake evolution.
Finally, the influence of atmospheric stability and inter-farm spacing on interactions within wind-farm clusters is investigated. Farm-to-farm wake dynamics result from the interplay between turbulence-mediated recovery and gravity-wave-induced large-scale atmospheric adjustments. Reduced ambient mixing can favor long-range wake persistence, whereas turbulence generated by the upstream farm accelerates downstream farm wake recovery. Additionally, turbulence levels in the downstream farm rapidly equilibrate with those upstream. Under strongly stratified conditions, coherent gravity-wave-like vertical
