Abstract
To date, one of the main challenges and requirements in wave energy technologies is to design energy-maximising control for a wave energy converter (WEC) device to achieve the energy maximization production, so as to reduce the levelized cost of energy (LCoE). Hence, this study starts from the numerical modelling of a 1:20 scaled Wavestar-prototype device based on the open source WEC-SIMulator (WECSim) benchmark, which is developed by the National Renewable Energy Laboratory (NREL) and Sandia National Laboratories (Sandia).
Next, a hierarchical tracking structure is selected as the core idea of this study and it contains two different parts. The first high-level part includes the wave excitation moment (WEM) estimation and the determination of the optimal reference signal. Four simple but effective robust methods are considered to design WEM estimators using some practical ways with low computation complexity, such as (i) Unknown Input Observer (UIO) with linear matrix inequality (LMI), (ii) Luenberger observer (LO) with LMI, (iii) LO with pole-placement, and (iv) Adaptive sliding mode observer (ASMO). On the other hand, the Extended Kalman Filter (EKF) can perform well in estimating the instantaneous amplitude and frequency of WEM for optimal reference calculation.
The second low-level part is to design energy-maximising controller. For example, a mixed LQR/H∞ control and sliding mode control (SMC) based on model-following tracking strategy (continuous-time modelling), model-predictive control (MPC) velocity tracking with Gaussian Process (GP) models (discrete-time modelling) are proposed, respectively. The designed tracking system can provide a near-resonance operation for the PAWEC device to achieve maximizing energy capture. Finally, a comparison study is done between the MPC and robust control methods in order to give some discussions and analysis about the robustness and optimality of the PAWEC control system with and without added matched disturbance tests, in terms of absorbed energy, extracted energy, extracted power and power take-off (PTO) moment, etc.
The simulation results show that ASMO gives best performance demonstrating low estimation delay, owing to fast sliding mode response property. ASMO has strong robustness and best all-round performance. LO-PP method has the simplest structure and is very attractive in real applications. The mixed LQR/H∞ control and SMC methods can provide strong robustness for PAWEC system. But the main problem of robust Model-Following is large negative power in energy conversion. The approach of MPC tracking is one of the best methods in this PhD study. It can provide PAWEC system with good robustness and solve control input constraint properly.
The proposed novel tracking control methods (a) a robust Model-Following or (b) a robust MPC framework can help PAWEC system to achieve robustness enhancement and maintain m