Model Predictive Control Implementation for the Ocean Grazer Wave Energy Converter with a port-Hamiltonian Model

Abstract

The ocean is a perpetual source of energy that has being studied for the last decades in order to convert the energy from the waves and tides into electricity, through a clean process. So far, several devices have been designed for that purpose, including the Ocean Grazer, a novel energy harvesting device that will allow to extract up to 260 GWh per year and store up to 800 MWh from the waves, through a point absorber take-o system, ensuring a continuous supply to the grid. The system uses a novel multiple-piston multiple-pump (MP2) concept to maximize the extracted energy, however, a control strategy is required. In this project is presented a model predictive control (MPC) strategy with a port-Hamiltonian (pH) model using the open source language Python, with the advantage over other control strategies in the literature that doesn't require a wave prediction. Its open loop validation showed an acceptable accuracy when compared against a MATLAB counterpart, but taking considerable less computing time ( 28 times less). The control strategy was tested using a 2 1

oater array, resulting possible to obtain a piston con guration for the MP2 in few seconds, and guaranteeing an energy absorption with less than 5 % of error when compared with the theoretical maximum value.