CAREER: Efficient Experimental Optimization for High-Performance Airborne Wind Energy Systems

This Faculty Early Career Development (CAREER) grant will pioneer a first-in-world rapid prototyping system for experimentally optimizing the flight dynamics and control of airborne wind energy systems. Airborne wind energy systems replace towers with tethers and a lifting body, reducing deployment time and fixed infrastructure costs, and enabling turbines to take advantage of strong, high-altitude winds. Successful realization of these systems is projected to yield levelized costs of electricity below $0.25 per kW-h, providing cost-competitive energy solutions to remote communities, islands, military bases, and deep-water offshore locations. The synthesis of control systems to stabilize airborne wind energy systems in harsh atmospheric conditions remains a bottleneck for their widespread acceptance, further exacerbated by high prototype development costs. This research will reduce control system prototyping costs by multiple orders of magnitude, using 1/100-scale models that are 3D printed, tethered, and 'flown' in a water channel laboratory test facility. The water channel provides an ideal mechanism for optimizing the control system design while replicating key dynamic properties of the full-scale system. Throughout the project, students will develop an industrial and small-business perspective through interactions with a leading early-stage airborne wind energy company. Outreach activities include the development of kite design modules for a high school engineering summer camp and co-design of an energy-rich science curriculum for an early college high school for economically disadvantaged students.

Optimization of airborne wind energy flight performance represents a coupled plant and controller optimization problem, where experiments are indispensable but expensive at full-scale. This research addresses the plant/controller coupling and the necessity of experiments through the a unique framework that combines numerical optimization with lab-scale experiments on 3D printed models that are tethered and 'flown' in a water channel. This water channel platform, which will be instrumented for closed-loop control of tethered systems, has been shown to yield provably similar dynamic performance to full-scale systems. In the proposed plant and controller optimization process, experimental data will be used to perform parameter identification and generate corrections to subsequent numerical optimization iterations. At the completion of each numerical optimization iteration, optimal design of experiments techniques will be used to determine a set of configurations to be tested, taking into account the cost of each reconfiguration. The research will focus on the derivation of convergence and efficiency results for the proposed algorithms, leveraging tools from system identification and optimal design of experiments. Furthermore, the optimization methods originating from this research will be validated on both a stationary and crosswind airborne wind energy system.