Simulation of Airborne Wind Energy Systems in the Atmospheric Boundary Layer
Airborne wind energy is the umbrella appellation for tethered airborne devices harvesting energy from high altitude wind resources. This emerging technology is the breeding ground for a diverse ecosystem of companies and research institutes developing numerous system archetypes and prototypes. The wakes of airborne wind energy systems (AWES) are commonly neglected for the small-scale prototypes currently under development. Consequently, little or none is known about the shape, the strength, or the size of AWES wakes. In particular, for the future utility-scale deployment of large-scale systems in farms, the adverse impact of wakes may play a significant role in the economic viability of the technology. The main goal of this dissertation is to investigate the wakes of airborne wind energy systems and assess their eventual impact on power performance of large airborne wind energy farms.
In the current dissertation, the virtual wind environment is computed by means of large-eddy simulations using the SP–Wind research code developed at KU Leuven, Belgium. The capability of the research code is extended to account for the presence of AWESs and their interaction with the wind flow. Within SP–Wind, the wind environment is modelled as pressure-driven boundary layer and coupled to an AWES flight simulator through actuator methods. The computational framework is ultimately combined with the modelling and optimal control toolbox awebox developed at the University of Freiburg, Germany. The capability of the simulation framework is first assessed for simplified configurations resembling the operation of conventional wind turbines and AWESs using predefined trajectories. In two initial studies, significant wakes of annular shape are identified. The unique characteristics of crosswind flight engender complex flow structures and interactions with the wind flow. Whereas for conventional wind turbines and AWESs with on-board power generation wakes are continuously induced, wake generation of pumping-mode AWESs with ground-based power generation is intermittent. Wakes are first induced during the power-generating phase, associated to the high wing loading of crosswind flight, and subsequently interrupted during the retraction phase due to the reduced wing loading.
In a next step, a nonlinear model predictive controller is added to the framework. This addition paves the road towards fully-coupled simulations of airborne wind energy farms. In preparation of the farm simulations, detailed investigations of anticipated future utility-scale AWESs, using both ground-based and on-board power generation, are conducted. The investigations of the pumping-mode AWES have also revealed the possible interaction of the tethered wing with its own wake. It is shown that restricting the reel-out speed of the tether according to an induction-based limitation can reduce substantially the wing–wake interaction.
Finally, wake interaction and power performance are addressed in the context of airborne wind energy farms where three park configurations are considered. The different park configurations consist of 25 systems operating in fully turbulent inflow conditions provided by concurrent precursor simulations. The first park configuration consists of pumping-mode AWESs arranged in a moderate density farm layout. The two other park configurations consist of on-board generation AWESs arranged in moderate and high density farm layouts. The computational framework combining large-eddy simulations and optimal control techniques captures reasonably well the complex interaction between the airborne wind energy farm and the virtual wind environment. The combined effects of wake and tracking-induced power losses specific to the investigated park configurations lead to farm efficiencies of approximately 82% for the park of pumping-mode systems and respectively 89% and 75% for the moderate and high density farm layouts of on-board generation systems.