Designing Renewables for an Uncertain World: Adaptive Co-design Formulation and Marine Energy Case Study

Renewable energy systems are subject to two key sources of uncertainty: the irreducible uncertainty associated with harvesting energy from a stochastic environmental resource and modeling uncertainty, which can decay over time. To maximize the energetic performance of a system subject to uncertainty, real-time plant and controller adaptability can be introduced. Real-time adaptation enables modification of aspects of a system's design (physical plant parameters through morphing or controller parameters through adaptive control) during operation. While the inclusion of plant and controller adaptability enhances energetic performance, adaptability comes at an economic cost, necessitating additional actuators and more advanced sensors. In this work, a co-design framework is presented to identify the system design, experimental characterization technique, and degree of plant and controller adaptability that minimize the expected levelized cost of energy (LCOE) of a renewable energy system. This framework is applied to a case study involving an installation of marine hydrokinetic kites using a detailed cost model and a computationally efficient surrogate model of energetic performance. Here, the optimal wing geometry, tether length, range of chord length (which can be adjusted during operation), and flight control architecture were identified. The expected LCOE-optimal degree of model refinement was also identified to account for uncertainty in the tether drag coefficient. To leverage the optimized systems range of adaptation, a real-time extremum-seeking-based adaptation strategy was developed. Dynamic simulations were then conducted wherein the energetic performance of the kite converged to the predictions generated by the surrogate model.