The impact of modelling and prediction errors on the performance of optimally controlled multi-DOF wave energy converters

Abstract

A well-conceived real-time control strategy can greatly increase the captured power for a wave energy converter (WEC). Optimal strategies rely on a dynamic model of the WEC and prediction of the wave excitation force several seconds into the future. Both the modelling and prediction processes are subject to errors. This paper investigates the impact of these errors on the performance of a multi-DOF submerged point absorber WEC. A state-space model of the system in its nominal position is derived and used by the control strategy. This idealised system is tested in multiple numerically generated irregular sea states with perfect estimation and prediction of the excitation force assumed. An optimally tuned passively damped system is used as a performance benchmark. The idealised system under optimal control is capable of more than doubling the captured power compared to the passively damped system. The control strategy is then applied to a full kinematic model of the WEC in the WEC-Sim environment. Real-time estimation and prediction of the excitation forces and constraints on motion and control force are also included. Under these more realistic conditions, the power gain is a more modest 68% at best across the tested sea states, and for one tested sea state there is no power gain compared to the passive system. Overall the gains are still significant and demonstrate the potential benefits of such control strategies for application to multi-DOF WECs, though more robust alternatives may be preferable.