Toward defining wave prediction requirements for wave energy applications

Abstract

Maximising the electrical output of wave energy converters through advanced control strategies depends significantly on the accuracy of incoming wave predictions. Various wave prediction techniques have been developed specifically for wave energy applications, typically operating under the assumption that greater accuracy leads to better performance. However, a clear understanding of how prediction errors influence controller performance remains lacking. This study explores the impact of wave prediction accuracy on energy output, aiming to define performance benchmarks for wave prediction models. Three hypothetical prediction error sources have been investigated, including the time delay, restricted frequency bandwidth in the predicted wave signal, and the prediction noise. The study was done specifically for the CETO wave energy converter and the test site located in Spain. The results indicate that to keep power losses due to wave prediction errors below 5\%, the prediction accuracy should achieve a normalised root mean square error (NRMSE) of less than 9\%.