Control of a point absorber wave energy converter in extreme wave conditions using a deep learning model in WEC-Sim

The survivability of wave energy converters (WECs) is one of the challenges that have a direct influence on their cost. To protect the WEC from the impact of extreme waves, it is often to over-dimension the components or adopt survivability modes e.g. by submerging or lifting the WEC if it is applicable. Here, a control strategy for adjusting the system damping is developed based on deep neural networks (DNN) to minimize the line (mooring) force exerted on a 1:30 scaled WEC. This DNN model is then implemented in a control system of a numerical WEC-Sim model to find the optimal power take-off (PTO) damping for every zero up-crossing episode of surface elevation which minimizes the peak line force. The WEC-Sim model was calibrated based on a 1:30 scaled wave tank experiment that was designed to investigate the WEC response in extreme sea states with a 50-year return period. It is shown that this survival strategy reduces the peak forces when compared with the response of a system that has been set to a constant PTO damping for the entire duration of the sea state.