Abstract
The Cycloidal Wave Energy Converter (CycWEC) is a device in which two hydrofoils rotate about a center shaft interacting with the water surface through the generation of circulatory forces. The CycWEC is a unique device as it classified as a single-sided, wave terminator by its ability to generate a single sided wave, but the device requires real-time estimation and control to efficiently extract energy from incoming spatio-temporal varying wave fields. Specifically, the CycWEC requires spatial alignment to the incoming wave crest, phase alignment of the main shaft with the passing wave packet, and proper pitch angle to wave height ratio to achieve optimal energy extraction. To account for these conditions, the wave state at the location of the wave energy converter must be fully observable in real time. This paper makes use of up-wave, noisy, velocity measurements from an acoustic Doppler velocimeter (ADP) to predict the wave state (phase, frequency, amplitude) at the location of the CycWEC. These predictions are then used to control the CycWEC in a variety of realistic sea conditions. The results indicate that autoregressive exogenous models with proper spatial and temporal filtering are suitable for robust predictions of current and future states. Specifically, proper orthogonal decomposition is used to identify an Airy basis function restricting ADP measurements to a physics based sub-domain. In addition, the dynamics of the CycWEC are represented by a quasithree-dimensional numerical method which leverages a low order potential flow model and free surface boundary conditions. The entire radiation code is integrated within a control volume in order to compute energy flux and ultimately capture width length. The impact of the realtime estimation scheme is contrasted with ideal control parameters, and the influence of the estimation error is shown to be negligible to the CycWEC efficiency.