Co-Design for Real-Time Adaptability: Methodology, Implementation, and Real-Time Performance on a Morphing Wind Turbine

Large-scale wind energy-harvesting systems operate in variable environments and are subject to long design and manufacturing cycles that often mandate design freezes before the system has been modeled with high certainty. Both challenges can be addressed by incorporating real-time adaptability into the plant and controller, though this comes at a cost. The proposed co-design framework aims to maximize the lifetime expected profit of a wind energy system (through the mathematical underpinnings are generalizable to other energy-harvesting devices) by considering a low-complexity surrogate model, a statistical model of the environment, a statistical characterization of how modeling uncertainty diminishes over the design cycle, and models describing the cost of adaptability and engineering labor. By jointly considering plant and controller adaptability, the framework balances high-cost/high-reward plant adaptability with lower-cost/lower-reward controller adaptability. This co-design framework is coupled with an online control strategy to enable real-time adaptation and applied to the segmented ultralight morphing rotor (SUMR) wind turbine to evaluate this approach. Applying this framework to the SUMR, and using the aforementioned surrogate model, the expected lifetime profit is shown to increase by 16.7% when the level of adaptability is optimized. Using the optimized design, a 24-hour dynamic simulation compared the performance of an adaptive control strategy to a non-adaptive strategy, showing a 4.4% increase in energy generation with the adaptive approach. These results closely align with predictions from the surrogate model, validating the co-design approach.