On Optimal Control for Nonlinear Hydraulic PTO in Wave Energy Conversion

The control system is the heart of a Wave Energy Conveter (WEC) capable to enhance the performance of the device and, along with the a proper techno-economic design, is in charge of the maximization of the converter performances. Optimal control approaches are (usually) employed to enhance the energy extraction performances of a WEC with the aim to improve the annual production. This work proposes a nonlinear Model Predictive Control (MPC) applied to the ISWEC, a gyroscopic-based WEC developed at the Politecnico di Torino (Italy), equipped with a Hydrauli Power Take-Off (HPTO). First, the Wave Excitation Force (WEF) estimation is addressed with a Neural Network (NN) model (model-free approach). The estimators is tuned at the operating conditions of the ISWEC in Pantelleria climate, a sea site in the Mediterranean Sea (Italy). The NN is designed with respect to the number of neurons and the delay step of the signals in input. Then, an optimal cotrol command is derived through the Pontryagin’s maximum principle to control the HPTO with a declutching logic in a optimal way. The formulation is reduced to a bang-bang type control that gives the optimal switching sequence. The NN is considered valid to address the WEF estimation or any other unknown input estimation for the ISWEC, leading for the DoF of interest an average accuracy always higher than 94%. Moreover, annual energy production of the ISWEC device is computed with the MPC control and compared with the original declutching control logic. Promising results are obtained since an increase of 6.5% on average of the annual energy production.